Date: 2019-12-26 00:12:13 CET, cola version: 1.3.2
Document is loading...
All available functions which can be applied to this res_list object:
res_list
#> A 'ConsensusPartitionList' object with 24 methods.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows are extracted by 'SD, CV, MAD, ATC' methods.
#> Subgroups are detected by 'hclust, kmeans, skmeans, pam, mclust, NMF' method.
#> Number of partitions are tried for k = 2, 3, 4, 5, 6.
#> Performed in total 30000 partitions by row resampling.
#>
#> Following methods can be applied to this 'ConsensusPartitionList' object:
#> [1] "cola_report" "collect_classes" "collect_plots" "collect_stats"
#> [5] "colnames" "functional_enrichment" "get_anno_col" "get_anno"
#> [9] "get_classes" "get_matrix" "get_membership" "get_stats"
#> [13] "is_best_k" "is_stable_k" "ncol" "nrow"
#> [17] "rownames" "show" "suggest_best_k" "test_to_known_factors"
#> [21] "top_rows_heatmap" "top_rows_overlap"
#>
#> You can get result for a single method by, e.g. object["SD", "hclust"] or object["SD:hclust"]
#> or a subset of methods by object[c("SD", "CV")], c("hclust", "kmeans")]
The call of run_all_consensus_partition_methods() was:
#> run_all_consensus_partition_methods(data = mat, mc.cores = 4)
Dimension of the input matrix:
mat = get_matrix(res_list)
dim(mat)
#> [1] 16571 54
The density distribution for each sample is visualized as in one column in the following heatmap. The clustering is based on the distance which is the Kolmogorov-Smirnov statistic between two distributions.
library(ComplexHeatmap)
densityHeatmap(mat, ylab = "value", cluster_columns = TRUE, show_column_names = FALSE,
mc.cores = 4)
Folowing table shows the best k (number of partitions) for each combination
of top-value methods and partition methods. Clicking on the method name in
the table goes to the section for a single combination of methods.
The cola vignette explains the definition of the metrics used for determining the best number of partitions.
suggest_best_k(res_list)
| The best k | 1-PAC | Mean silhouette | Concordance | Optional k | ||
|---|---|---|---|---|---|---|
| SD:kmeans | 2 | 1.000 | 0.995 | 0.998 | ** | |
| SD:pam | 6 | 1.000 | 0.974 | 0.989 | ** | 3,5 |
| MAD:kmeans | 2 | 1.000 | 0.983 | 0.990 | ** | |
| MAD:skmeans | 2 | 1.000 | 0.967 | 0.986 | ** | |
| ATC:hclust | 2 | 1.000 | 0.961 | 0.983 | ** | |
| ATC:kmeans | 2 | 1.000 | 0.975 | 0.989 | ** | |
| ATC:NMF | 2 | 1.000 | 0.982 | 0.992 | ** | |
| SD:NMF | 2 | 0.999 | 0.962 | 0.984 | ** | |
| MAD:pam | 6 | 0.970 | 0.896 | 0.961 | ** | |
| ATC:skmeans | 4 | 0.929 | 0.903 | 0.959 | * | 2,3 |
| SD:skmeans | 2 | 0.923 | 0.937 | 0.973 | * | |
| ATC:pam | 6 | 0.908 | 0.829 | 0.937 | * | 3,4 |
| CV:pam | 6 | 0.905 | 0.913 | 0.958 | * | 2,3,5 |
| MAD:hclust | 2 | 0.902 | 0.958 | 0.975 | * | |
| MAD:NMF | 2 | 0.885 | 0.932 | 0.971 | ||
| CV:skmeans | 3 | 0.882 | 0.894 | 0.954 | ||
| CV:NMF | 3 | 0.875 | 0.876 | 0.950 | ||
| SD:mclust | 2 | 0.759 | 0.950 | 0.965 | ||
| ATC:mclust | 2 | 0.752 | 0.916 | 0.962 | ||
| CV:mclust | 2 | 0.747 | 0.858 | 0.939 | ||
| CV:hclust | 3 | 0.656 | 0.773 | 0.910 | ||
| CV:kmeans | 2 | 0.549 | 0.882 | 0.927 | ||
| MAD:mclust | 3 | 0.504 | 0.743 | 0.862 | ||
| SD:hclust | NA | NA | NA | NA |
**: 1-PAC > 0.95, *: 1-PAC > 0.9
Cumulative distribution function curves of consensus matrix for all methods.
collect_plots(res_list, fun = plot_ecdf)
Consensus heatmaps for all methods. (What is a consensus heatmap?)
collect_plots(res_list, k = 2, fun = consensus_heatmap, mc.cores = 4)
collect_plots(res_list, k = 3, fun = consensus_heatmap, mc.cores = 4)
collect_plots(res_list, k = 4, fun = consensus_heatmap, mc.cores = 4)
collect_plots(res_list, k = 5, fun = consensus_heatmap, mc.cores = 4)
collect_plots(res_list, k = 6, fun = consensus_heatmap, mc.cores = 4)
Membership heatmaps for all methods. (What is a membership heatmap?)
collect_plots(res_list, k = 2, fun = membership_heatmap, mc.cores = 4)
collect_plots(res_list, k = 3, fun = membership_heatmap, mc.cores = 4)
collect_plots(res_list, k = 4, fun = membership_heatmap, mc.cores = 4)
collect_plots(res_list, k = 5, fun = membership_heatmap, mc.cores = 4)
collect_plots(res_list, k = 6, fun = membership_heatmap, mc.cores = 4)
Signature heatmaps for all methods. (What is a signature heatmap?)
Note in following heatmaps, rows are scaled.
collect_plots(res_list, k = 2, fun = get_signatures, mc.cores = 4)
collect_plots(res_list, k = 3, fun = get_signatures, mc.cores = 4)
collect_plots(res_list, k = 4, fun = get_signatures, mc.cores = 4)
collect_plots(res_list, k = 5, fun = get_signatures, mc.cores = 4)
collect_plots(res_list, k = 6, fun = get_signatures, mc.cores = 4)
The statistics used for measuring the stability of consensus partitioning. (How are they defined?)
get_stats(res_list, k = 2)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> SD:NMF 2 0.999 0.962 0.984 0.485 0.516 0.516
#> CV:NMF 2 0.641 0.847 0.926 0.480 0.508 0.508
#> MAD:NMF 2 0.885 0.932 0.971 0.491 0.508 0.508
#> ATC:NMF 2 1.000 0.982 0.992 0.463 0.535 0.535
#> SD:skmeans 2 0.923 0.937 0.973 0.502 0.502 0.502
#> CV:skmeans 2 0.849 0.881 0.955 0.503 0.497 0.497
#> MAD:skmeans 2 1.000 0.967 0.986 0.509 0.491 0.491
#> ATC:skmeans 2 1.000 0.957 0.984 0.507 0.497 0.497
#> SD:mclust 2 0.759 0.950 0.965 0.450 0.525 0.525
#> CV:mclust 2 0.747 0.858 0.939 0.440 0.575 0.575
#> MAD:mclust 2 0.520 0.814 0.909 0.356 0.693 0.693
#> ATC:mclust 2 0.752 0.916 0.962 0.470 0.535 0.535
#> SD:kmeans 2 1.000 0.995 0.998 0.474 0.525 0.525
#> CV:kmeans 2 0.549 0.882 0.927 0.434 0.535 0.535
#> MAD:kmeans 2 1.000 0.983 0.990 0.478 0.525 0.525
#> ATC:kmeans 2 1.000 0.975 0.989 0.474 0.525 0.525
#> SD:pam 2 0.820 0.892 0.953 0.415 0.609 0.609
#> CV:pam 2 1.000 0.958 0.982 0.352 0.628 0.628
#> MAD:pam 2 0.887 0.939 0.974 0.502 0.497 0.497
#> ATC:pam 2 0.823 0.888 0.957 0.480 0.525 0.525
#> SD:hclust 2 0.454 0.924 0.899 0.346 0.560 0.560
#> CV:hclust 2 0.398 0.743 0.859 0.292 0.669 0.669
#> MAD:hclust 2 0.902 0.958 0.975 0.437 0.560 0.560
#> ATC:hclust 2 1.000 0.961 0.983 0.438 0.575 0.575
get_stats(res_list, k = 3)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> SD:NMF 3 0.893 0.901 0.960 0.2981 0.790 0.619
#> CV:NMF 3 0.875 0.876 0.950 0.3501 0.735 0.527
#> MAD:NMF 3 0.662 0.684 0.833 0.2903 0.778 0.596
#> ATC:NMF 3 0.638 0.702 0.849 0.1773 0.913 0.844
#> SD:skmeans 3 0.894 0.924 0.965 0.3182 0.718 0.496
#> CV:skmeans 3 0.882 0.894 0.954 0.3124 0.734 0.515
#> MAD:skmeans 3 0.673 0.659 0.827 0.2770 0.837 0.678
#> ATC:skmeans 3 0.976 0.951 0.975 0.2101 0.875 0.752
#> SD:mclust 3 0.430 0.467 0.719 0.3199 0.709 0.490
#> CV:mclust 3 0.346 0.248 0.678 0.3211 0.640 0.476
#> MAD:mclust 3 0.504 0.743 0.862 0.4658 0.756 0.655
#> ATC:mclust 3 0.517 0.712 0.858 0.2427 0.755 0.603
#> SD:kmeans 3 0.626 0.775 0.886 0.2425 0.901 0.814
#> CV:kmeans 3 0.658 0.828 0.905 0.3443 0.855 0.737
#> MAD:kmeans 3 0.517 0.249 0.660 0.3028 0.900 0.817
#> ATC:kmeans 3 0.829 0.889 0.926 0.3837 0.737 0.529
#> SD:pam 3 0.911 0.876 0.936 0.2619 0.844 0.748
#> CV:pam 3 0.968 0.928 0.978 0.3136 0.876 0.804
#> MAD:pam 3 0.719 0.697 0.867 0.2247 0.860 0.732
#> ATC:pam 3 0.929 0.913 0.966 0.4073 0.739 0.530
#> SD:hclust 3 0.832 0.887 0.921 0.4217 0.951 0.913
#> CV:hclust 3 0.656 0.773 0.910 0.5074 0.915 0.874
#> MAD:hclust 3 0.923 0.932 0.980 0.0447 0.989 0.980
#> ATC:hclust 3 0.678 0.831 0.840 0.2002 0.980 0.966
get_stats(res_list, k = 4)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> SD:NMF 4 0.737 0.682 0.816 0.1612 0.771 0.472
#> CV:NMF 4 0.632 0.455 0.716 0.1236 0.881 0.682
#> MAD:NMF 4 0.813 0.891 0.934 0.1585 0.783 0.493
#> ATC:NMF 4 0.599 0.790 0.837 0.2642 0.732 0.493
#> SD:skmeans 4 0.853 0.868 0.936 0.1223 0.866 0.627
#> CV:skmeans 4 0.721 0.652 0.851 0.1098 0.881 0.670
#> MAD:skmeans 4 0.824 0.861 0.939 0.1388 0.820 0.544
#> ATC:skmeans 4 0.929 0.903 0.959 0.1219 0.906 0.762
#> SD:mclust 4 0.473 0.706 0.767 0.1476 0.778 0.504
#> CV:mclust 4 0.465 0.533 0.715 0.1654 0.748 0.531
#> MAD:mclust 4 0.471 0.597 0.805 0.2518 0.903 0.805
#> ATC:mclust 4 0.533 0.492 0.739 0.1802 0.813 0.597
#> SD:kmeans 4 0.580 0.716 0.808 0.1788 0.809 0.573
#> CV:kmeans 4 0.619 0.640 0.821 0.1648 0.886 0.743
#> MAD:kmeans 4 0.612 0.657 0.791 0.1507 0.708 0.434
#> ATC:kmeans 4 0.722 0.699 0.822 0.1019 0.960 0.883
#> SD:pam 4 0.831 0.846 0.921 0.1367 0.957 0.910
#> CV:pam 4 0.966 0.927 0.977 0.0714 0.983 0.967
#> MAD:pam 4 0.891 0.784 0.895 0.1724 0.809 0.563
#> ATC:pam 4 0.985 0.921 0.968 0.0842 0.916 0.750
#> SD:hclust 4 0.871 0.899 0.965 0.0397 0.974 0.950
#> CV:hclust 4 0.595 0.654 0.851 0.1852 0.876 0.796
#> MAD:hclust 4 0.812 0.898 0.929 0.1296 0.956 0.920
#> ATC:hclust 4 0.704 0.670 0.846 0.2523 0.754 0.557
get_stats(res_list, k = 5)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> SD:NMF 5 0.707 0.701 0.811 0.0531 0.934 0.766
#> CV:NMF 5 0.760 0.725 0.826 0.0718 0.796 0.412
#> MAD:NMF 5 0.835 0.807 0.887 0.0476 0.963 0.864
#> ATC:NMF 5 0.653 0.700 0.841 0.0641 0.993 0.976
#> SD:skmeans 5 0.835 0.832 0.913 0.0627 0.920 0.705
#> CV:skmeans 5 0.733 0.713 0.834 0.0584 0.890 0.625
#> MAD:skmeans 5 0.882 0.846 0.925 0.0659 0.911 0.676
#> ATC:skmeans 5 0.832 0.828 0.860 0.0702 0.916 0.737
#> SD:mclust 5 0.608 0.402 0.667 0.1466 0.845 0.574
#> CV:mclust 5 0.529 0.541 0.689 0.1093 0.646 0.258
#> MAD:mclust 5 0.674 0.674 0.818 0.2036 0.724 0.382
#> ATC:mclust 5 0.539 0.474 0.702 0.0914 0.742 0.380
#> SD:kmeans 5 0.668 0.672 0.788 0.0841 0.932 0.761
#> CV:kmeans 5 0.663 0.572 0.775 0.0763 0.893 0.707
#> MAD:kmeans 5 0.767 0.748 0.837 0.0761 0.867 0.563
#> ATC:kmeans 5 0.724 0.676 0.793 0.0640 0.904 0.703
#> SD:pam 5 1.000 0.948 0.980 0.1743 0.822 0.612
#> CV:pam 5 0.942 0.902 0.969 0.2972 0.845 0.691
#> MAD:pam 5 0.848 0.792 0.916 0.0250 0.785 0.425
#> ATC:pam 5 0.857 0.792 0.919 0.0692 0.915 0.697
#> SD:hclust 5 0.868 0.837 0.938 0.0311 0.994 0.989
#> CV:hclust 5 0.604 0.667 0.874 0.0894 0.904 0.811
#> MAD:hclust 5 0.661 0.771 0.909 0.0968 0.922 0.847
#> ATC:hclust 5 0.663 0.705 0.846 0.0175 0.904 0.730
get_stats(res_list, k = 6)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> SD:NMF 6 0.835 0.776 0.892 0.0475 0.908 0.633
#> CV:NMF 6 0.880 0.828 0.916 0.0468 0.952 0.780
#> MAD:NMF 6 0.816 0.733 0.869 0.0396 0.954 0.812
#> ATC:NMF 6 0.647 0.626 0.815 0.0411 0.844 0.549
#> SD:skmeans 6 0.814 0.808 0.888 0.0308 0.962 0.824
#> CV:skmeans 6 0.792 0.741 0.853 0.0376 0.969 0.852
#> MAD:skmeans 6 0.862 0.783 0.890 0.0315 0.945 0.759
#> ATC:skmeans 6 0.817 0.776 0.861 0.0339 0.980 0.923
#> SD:mclust 6 0.712 0.607 0.745 0.0255 0.847 0.454
#> CV:mclust 6 0.627 0.538 0.708 0.0601 0.843 0.416
#> MAD:mclust 6 0.701 0.788 0.831 0.0373 0.913 0.607
#> ATC:mclust 6 0.778 0.806 0.837 0.0753 0.840 0.475
#> SD:kmeans 6 0.789 0.634 0.810 0.0512 0.959 0.835
#> CV:kmeans 6 0.707 0.508 0.767 0.0567 0.920 0.732
#> MAD:kmeans 6 0.809 0.687 0.806 0.0443 0.957 0.801
#> ATC:kmeans 6 0.738 0.683 0.783 0.0524 0.891 0.574
#> SD:pam 6 1.000 0.974 0.989 0.0487 0.957 0.859
#> CV:pam 6 0.905 0.913 0.958 0.0688 0.957 0.877
#> MAD:pam 6 0.970 0.896 0.961 0.0716 0.888 0.608
#> ATC:pam 6 0.908 0.829 0.937 0.0202 0.963 0.831
#> SD:hclust 6 0.811 0.842 0.933 0.0486 0.990 0.979
#> CV:hclust 6 0.685 0.668 0.845 0.0525 0.933 0.860
#> MAD:hclust 6 0.721 0.819 0.915 0.1253 0.949 0.883
#> ATC:hclust 6 0.693 0.568 0.825 0.0614 0.881 0.652
Following heatmap plots the partition for each combination of methods and the lightness correspond to the silhouette scores for samples in each method. On top the consensus subgroup is inferred from all methods by taking the mean silhouette scores as weight.
collect_stats(res_list, k = 2)
collect_stats(res_list, k = 3)
collect_stats(res_list, k = 4)
collect_stats(res_list, k = 5)
collect_stats(res_list, k = 6)
Collect partitions from all methods:
collect_classes(res_list, k = 2)
collect_classes(res_list, k = 3)
collect_classes(res_list, k = 4)
collect_classes(res_list, k = 5)
collect_classes(res_list, k = 6)
Overlap of top rows from different top-row methods:
top_rows_overlap(res_list, top_n = 1000, method = "euler")
top_rows_overlap(res_list, top_n = 2000, method = "euler")
top_rows_overlap(res_list, top_n = 3000, method = "euler")
top_rows_overlap(res_list, top_n = 4000, method = "euler")
top_rows_overlap(res_list, top_n = 5000, method = "euler")
Also visualize the correspondance of rankings between different top-row methods:
top_rows_overlap(res_list, top_n = 1000, method = "correspondance")
top_rows_overlap(res_list, top_n = 2000, method = "correspondance")
top_rows_overlap(res_list, top_n = 3000, method = "correspondance")
top_rows_overlap(res_list, top_n = 4000, method = "correspondance")
top_rows_overlap(res_list, top_n = 5000, method = "correspondance")
Heatmaps of the top rows:
top_rows_heatmap(res_list, top_n = 1000)
top_rows_heatmap(res_list, top_n = 2000)
top_rows_heatmap(res_list, top_n = 3000)
top_rows_heatmap(res_list, top_n = 4000)
top_rows_heatmap(res_list, top_n = 5000)
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "hclust"]
# you can also extract it by
# res = res_list["SD:hclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'hclust' method.
#> Performed in total 1250 partitions by row resampling.
#> There is no best k.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.454 0.924 0.899 0.3464 0.560 0.560
#> 3 3 0.832 0.887 0.921 0.4217 0.951 0.913
#> 4 4 0.871 0.899 0.965 0.0397 0.974 0.950
#> 5 5 0.868 0.837 0.938 0.0311 0.994 0.989
#> 6 6 0.811 0.842 0.933 0.0486 0.990 0.979
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] NA
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.965 0.000 1.000
#> SRR1274529 2 0.000 0.965 0.000 1.000
#> SRR1274530 2 0.000 0.965 0.000 1.000
#> SRR1274531 2 0.327 0.909 0.060 0.940
#> SRR1274532 2 0.000 0.965 0.000 1.000
#> SRR1274533 2 0.000 0.965 0.000 1.000
#> SRR1274534 2 0.000 0.965 0.000 1.000
#> SRR1274535 2 0.000 0.965 0.000 1.000
#> SRR1274536 2 0.402 0.883 0.080 0.920
#> SRR1274537 2 0.000 0.965 0.000 1.000
#> SRR1274538 2 0.000 0.965 0.000 1.000
#> SRR1274539 2 0.000 0.965 0.000 1.000
#> SRR1274540 2 0.000 0.965 0.000 1.000
#> SRR1274541 2 0.000 0.965 0.000 1.000
#> SRR1274542 2 0.000 0.965 0.000 1.000
#> SRR1274543 2 0.000 0.965 0.000 1.000
#> SRR1274544 2 0.000 0.965 0.000 1.000
#> SRR1274545 2 0.000 0.965 0.000 1.000
#> SRR1274546 2 0.000 0.965 0.000 1.000
#> SRR1274547 2 0.000 0.965 0.000 1.000
#> SRR1274548 2 0.000 0.965 0.000 1.000
#> SRR1274549 2 0.855 0.457 0.280 0.720
#> SRR1274550 1 0.833 0.926 0.736 0.264
#> SRR1274551 2 0.000 0.965 0.000 1.000
#> SRR1274552 1 0.833 0.926 0.736 0.264
#> SRR1274553 1 0.833 0.926 0.736 0.264
#> SRR1274554 2 0.000 0.965 0.000 1.000
#> SRR1274555 1 0.833 0.926 0.736 0.264
#> SRR1274556 1 0.833 0.926 0.736 0.264
#> SRR1274557 1 0.833 0.926 0.736 0.264
#> SRR1274558 2 0.506 0.834 0.112 0.888
#> SRR1274559 2 0.506 0.834 0.112 0.888
#> SRR1274560 2 0.506 0.834 0.112 0.888
#> SRR1274561 1 0.881 0.907 0.700 0.300
#> SRR1274562 1 0.814 0.928 0.748 0.252
#> SRR1274563 1 0.833 0.926 0.736 0.264
#> SRR1274564 2 0.402 0.883 0.080 0.920
#> SRR1274565 2 0.000 0.965 0.000 1.000
#> SRR1274566 2 0.141 0.949 0.020 0.980
#> SRR1274567 2 0.141 0.949 0.020 0.980
#> SRR1274568 2 0.278 0.921 0.048 0.952
#> SRR1274569 1 0.833 0.927 0.736 0.264
#> SRR1274570 1 0.833 0.927 0.736 0.264
#> SRR1274571 1 0.833 0.927 0.736 0.264
#> SRR1274572 1 0.833 0.927 0.736 0.264
#> SRR1274573 2 0.000 0.965 0.000 1.000
#> SRR1274574 2 0.000 0.965 0.000 1.000
#> SRR1274575 2 0.000 0.965 0.000 1.000
#> SRR1274576 2 0.000 0.965 0.000 1.000
#> SRR1274577 2 0.000 0.965 0.000 1.000
#> SRR1274578 1 0.833 0.927 0.736 0.264
#> SRR1274579 1 0.833 0.927 0.736 0.264
#> SRR1274580 1 0.833 0.927 0.736 0.264
#> SRR1274581 1 0.833 0.601 0.736 0.264
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274531 2 0.2947 0.9069 0.020 0.920 0.060
#> SRR1274532 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274533 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274534 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274535 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274536 2 0.2537 0.9115 0.000 0.920 0.080
#> SRR1274537 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274539 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274549 2 0.7798 0.4398 0.080 0.624 0.296
#> SRR1274550 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274551 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274552 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274553 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274554 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274555 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274556 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274557 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274558 2 0.3192 0.8756 0.112 0.888 0.000
#> SRR1274559 2 0.3192 0.8756 0.112 0.888 0.000
#> SRR1274560 2 0.3192 0.8756 0.112 0.888 0.000
#> SRR1274561 1 0.8013 0.0478 0.588 0.080 0.332
#> SRR1274562 1 0.5905 0.0354 0.648 0.000 0.352
#> SRR1274563 3 0.5397 1.0000 0.280 0.000 0.720
#> SRR1274564 2 0.2537 0.9115 0.000 0.920 0.080
#> SRR1274565 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274566 2 0.0892 0.9587 0.000 0.980 0.020
#> SRR1274567 2 0.0892 0.9587 0.000 0.980 0.020
#> SRR1274568 2 0.1753 0.9364 0.048 0.952 0.000
#> SRR1274569 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274577 2 0.0000 0.9703 0.000 1.000 0.000
#> SRR1274578 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.8038 1.000 0.000 0.000
#> SRR1274581 1 0.9050 0.3671 0.536 0.168 0.296
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274529 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274530 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274531 2 0.2011 0.902 0.000 0.920 0.080 0
#> SRR1274532 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274533 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274534 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274535 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274536 2 0.2011 0.907 0.000 0.920 0.080 0
#> SRR1274537 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274538 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274539 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274540 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274541 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274542 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274543 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274544 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274545 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274546 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274547 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274548 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274549 2 0.4776 0.430 0.000 0.624 0.376 0
#> SRR1274550 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274551 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274552 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274553 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274554 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274555 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274556 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274557 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274558 2 0.2530 0.871 0.112 0.888 0.000 0
#> SRR1274559 2 0.2530 0.871 0.112 0.888 0.000 0
#> SRR1274560 2 0.2530 0.871 0.112 0.888 0.000 0
#> SRR1274561 3 0.6453 0.364 0.360 0.080 0.560 0
#> SRR1274562 3 0.4855 0.382 0.400 0.000 0.600 0
#> SRR1274563 3 0.0000 0.860 0.000 0.000 1.000 0
#> SRR1274564 2 0.2011 0.907 0.000 0.920 0.080 0
#> SRR1274565 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274566 2 0.0707 0.957 0.000 0.980 0.020 0
#> SRR1274567 2 0.0707 0.957 0.000 0.980 0.020 0
#> SRR1274568 2 0.1389 0.934 0.048 0.952 0.000 0
#> SRR1274569 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274570 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274571 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274572 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274573 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274574 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274575 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274576 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274577 2 0.0000 0.969 0.000 1.000 0.000 0
#> SRR1274578 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274579 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274580 1 0.0000 1.000 1.000 0.000 0.000 0
#> SRR1274581 4 0.0000 0.000 0.000 0.000 0.000 1
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274530 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274531 2 0.1831 0.905 0.000 0.920 0.000 0.004 0.076
#> SRR1274532 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274534 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274535 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274536 2 0.2074 0.909 0.000 0.920 0.000 0.044 0.036
#> SRR1274537 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274540 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 2 0.6343 0.441 0.000 0.624 0.200 0.044 0.132
#> SRR1274550 3 0.4959 0.561 0.000 0.000 0.684 0.240 0.076
#> SRR1274551 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.4959 0.561 0.000 0.000 0.684 0.240 0.076
#> SRR1274553 3 0.4959 0.561 0.000 0.000 0.684 0.240 0.076
#> SRR1274554 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274555 3 0.0000 0.536 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 5 0.4171 0.000 0.000 0.000 0.396 0.000 0.604
#> SRR1274557 3 0.0000 0.536 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 2 0.3454 0.801 0.028 0.816 0.000 0.156 0.000
#> SRR1274559 2 0.3454 0.801 0.028 0.816 0.000 0.156 0.000
#> SRR1274560 2 0.3454 0.801 0.028 0.816 0.000 0.156 0.000
#> SRR1274561 3 0.5558 0.330 0.360 0.080 0.560 0.000 0.000
#> SRR1274562 3 0.4182 0.362 0.400 0.000 0.600 0.000 0.000
#> SRR1274563 3 0.0000 0.536 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 2 0.2074 0.909 0.000 0.920 0.000 0.044 0.036
#> SRR1274565 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274566 2 0.0693 0.954 0.000 0.980 0.000 0.012 0.008
#> SRR1274567 2 0.0693 0.954 0.000 0.980 0.000 0.012 0.008
#> SRR1274568 2 0.1701 0.921 0.016 0.936 0.000 0.048 0.000
#> SRR1274569 1 0.1792 0.891 0.916 0.000 0.000 0.084 0.000
#> SRR1274570 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274576 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 2 0.0000 0.965 0.000 1.000 0.000 0.000 0.000
#> SRR1274578 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.983 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 4 0.4171 0.000 0.000 0.000 0.000 0.604 0.396
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0858 0.932 0.000 0.968 0.000 0.028 0.004 0.000
#> SRR1274530 2 0.0858 0.932 0.000 0.968 0.000 0.028 0.004 0.000
#> SRR1274531 2 0.1967 0.886 0.000 0.904 0.000 0.084 0.012 0.000
#> SRR1274532 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274534 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274535 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274536 2 0.3215 0.834 0.000 0.828 0.000 0.072 0.100 0.000
#> SRR1274537 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274540 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0858 0.932 0.000 0.968 0.000 0.028 0.004 0.000
#> SRR1274546 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 2 0.6430 0.358 0.000 0.560 0.200 0.140 0.100 0.000
#> SRR1274550 5 0.1327 1.000 0.000 0.000 0.064 0.000 0.936 0.000
#> SRR1274551 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.1327 1.000 0.000 0.000 0.064 0.000 0.936 0.000
#> SRR1274553 5 0.1327 1.000 0.000 0.000 0.064 0.000 0.936 0.000
#> SRR1274554 2 0.0972 0.930 0.000 0.964 0.000 0.008 0.028 0.000
#> SRR1274555 3 0.0000 0.576 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 4 0.3684 0.000 0.000 0.000 0.332 0.664 0.004 0.000
#> SRR1274557 3 0.0000 0.576 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274558 2 0.3323 0.725 0.000 0.752 0.000 0.000 0.008 0.240
#> SRR1274559 2 0.3756 0.704 0.000 0.736 0.000 0.016 0.008 0.240
#> SRR1274560 2 0.3756 0.704 0.000 0.736 0.000 0.016 0.008 0.240
#> SRR1274561 3 0.5768 0.430 0.316 0.064 0.560 0.060 0.000 0.000
#> SRR1274562 3 0.4530 0.449 0.356 0.000 0.600 0.044 0.000 0.000
#> SRR1274563 3 0.0000 0.576 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 2 0.3215 0.834 0.000 0.828 0.000 0.072 0.100 0.000
#> SRR1274565 2 0.1049 0.928 0.000 0.960 0.000 0.008 0.032 0.000
#> SRR1274566 2 0.2258 0.892 0.000 0.896 0.000 0.044 0.060 0.000
#> SRR1274567 2 0.2258 0.892 0.000 0.896 0.000 0.044 0.060 0.000
#> SRR1274568 2 0.1327 0.908 0.000 0.936 0.000 0.000 0.000 0.064
#> SRR1274569 1 0.1957 0.851 0.888 0.000 0.000 0.000 0.000 0.112
#> SRR1274570 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.0858 0.932 0.000 0.968 0.000 0.028 0.004 0.000
#> SRR1274576 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 2 0.0000 0.943 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274578 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.977 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274581 6 0.3076 0.000 0.000 0.000 0.000 0.240 0.000 0.760
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "kmeans"]
# you can also extract it by
# res = res_list["SD:kmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'kmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.995 0.998 0.4744 0.525 0.525
#> 3 3 0.626 0.775 0.886 0.2425 0.901 0.814
#> 4 4 0.580 0.716 0.808 0.1788 0.809 0.573
#> 5 5 0.668 0.672 0.788 0.0841 0.932 0.761
#> 6 6 0.789 0.634 0.810 0.0512 0.959 0.835
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 1.000 0.000 1.000
#> SRR1274529 2 0.000 1.000 0.000 1.000
#> SRR1274530 2 0.000 1.000 0.000 1.000
#> SRR1274531 2 0.000 1.000 0.000 1.000
#> SRR1274532 2 0.000 1.000 0.000 1.000
#> SRR1274533 2 0.000 1.000 0.000 1.000
#> SRR1274534 2 0.000 1.000 0.000 1.000
#> SRR1274535 2 0.000 1.000 0.000 1.000
#> SRR1274536 2 0.000 1.000 0.000 1.000
#> SRR1274537 2 0.000 1.000 0.000 1.000
#> SRR1274538 2 0.000 1.000 0.000 1.000
#> SRR1274539 2 0.000 1.000 0.000 1.000
#> SRR1274540 2 0.000 1.000 0.000 1.000
#> SRR1274541 2 0.000 1.000 0.000 1.000
#> SRR1274542 2 0.000 1.000 0.000 1.000
#> SRR1274543 2 0.000 1.000 0.000 1.000
#> SRR1274544 2 0.000 1.000 0.000 1.000
#> SRR1274545 2 0.000 1.000 0.000 1.000
#> SRR1274546 2 0.000 1.000 0.000 1.000
#> SRR1274547 2 0.000 1.000 0.000 1.000
#> SRR1274548 2 0.000 1.000 0.000 1.000
#> SRR1274549 1 0.311 0.943 0.944 0.056
#> SRR1274550 1 0.000 0.993 1.000 0.000
#> SRR1274551 2 0.000 1.000 0.000 1.000
#> SRR1274552 1 0.358 0.931 0.932 0.068
#> SRR1274553 1 0.000 0.993 1.000 0.000
#> SRR1274554 2 0.000 1.000 0.000 1.000
#> SRR1274555 1 0.000 0.993 1.000 0.000
#> SRR1274556 1 0.000 0.993 1.000 0.000
#> SRR1274557 1 0.000 0.993 1.000 0.000
#> SRR1274558 2 0.000 1.000 0.000 1.000
#> SRR1274559 1 0.000 0.993 1.000 0.000
#> SRR1274560 1 0.000 0.993 1.000 0.000
#> SRR1274561 1 0.000 0.993 1.000 0.000
#> SRR1274562 1 0.000 0.993 1.000 0.000
#> SRR1274563 1 0.000 0.993 1.000 0.000
#> SRR1274564 2 0.000 1.000 0.000 1.000
#> SRR1274565 2 0.000 1.000 0.000 1.000
#> SRR1274566 2 0.000 1.000 0.000 1.000
#> SRR1274567 2 0.000 1.000 0.000 1.000
#> SRR1274568 2 0.000 1.000 0.000 1.000
#> SRR1274569 1 0.000 0.993 1.000 0.000
#> SRR1274570 1 0.000 0.993 1.000 0.000
#> SRR1274571 1 0.000 0.993 1.000 0.000
#> SRR1274572 1 0.000 0.993 1.000 0.000
#> SRR1274573 2 0.000 1.000 0.000 1.000
#> SRR1274574 2 0.000 1.000 0.000 1.000
#> SRR1274575 2 0.000 1.000 0.000 1.000
#> SRR1274576 2 0.000 1.000 0.000 1.000
#> SRR1274577 2 0.000 1.000 0.000 1.000
#> SRR1274578 1 0.000 0.993 1.000 0.000
#> SRR1274579 1 0.000 0.993 1.000 0.000
#> SRR1274580 1 0.000 0.993 1.000 0.000
#> SRR1274581 1 0.000 0.993 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274529 2 0.3551 0.8260 0.000 0.868 0.132
#> SRR1274530 2 0.3551 0.8260 0.000 0.868 0.132
#> SRR1274531 3 0.6309 -0.4177 0.000 0.496 0.504
#> SRR1274532 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274533 2 0.5706 0.5907 0.000 0.680 0.320
#> SRR1274534 2 0.4842 0.7217 0.000 0.776 0.224
#> SRR1274535 2 0.5905 0.5333 0.000 0.648 0.352
#> SRR1274536 2 0.6154 0.5366 0.000 0.592 0.408
#> SRR1274537 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274539 2 0.4842 0.7217 0.000 0.776 0.224
#> SRR1274540 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274545 2 0.1860 0.8649 0.000 0.948 0.052
#> SRR1274546 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.6897 0.000 0.000 1.000
#> SRR1274550 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274551 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274552 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274553 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274554 2 0.5835 0.5560 0.000 0.660 0.340
#> SRR1274555 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274556 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274557 3 0.3551 0.8035 0.132 0.000 0.868
#> SRR1274558 2 0.0592 0.8774 0.012 0.988 0.000
#> SRR1274559 1 0.3764 0.8296 0.892 0.040 0.068
#> SRR1274560 1 0.2845 0.8597 0.920 0.012 0.068
#> SRR1274561 1 0.5621 0.5555 0.692 0.000 0.308
#> SRR1274562 1 0.4654 0.7223 0.792 0.000 0.208
#> SRR1274563 3 0.3879 0.7799 0.152 0.000 0.848
#> SRR1274564 2 0.3879 0.8203 0.000 0.848 0.152
#> SRR1274565 2 0.3619 0.8252 0.000 0.864 0.136
#> SRR1274566 2 0.3879 0.8203 0.000 0.848 0.152
#> SRR1274567 2 0.6154 0.5366 0.000 0.592 0.408
#> SRR1274568 2 0.3207 0.8374 0.012 0.904 0.084
#> SRR1274569 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274575 2 0.3551 0.8260 0.000 0.868 0.132
#> SRR1274576 2 0.0000 0.8823 0.000 1.000 0.000
#> SRR1274577 2 0.5098 0.6955 0.000 0.752 0.248
#> SRR1274578 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.9145 1.000 0.000 0.000
#> SRR1274581 3 0.6309 0.0678 0.496 0.000 0.504
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.4222 0.5276 0.000 0.728 0.000 0.272
#> SRR1274530 2 0.4222 0.5276 0.000 0.728 0.000 0.272
#> SRR1274531 4 0.5678 0.5878 0.000 0.172 0.112 0.716
#> SRR1274532 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.7013 0.6943 0.000 0.356 0.128 0.516
#> SRR1274534 4 0.6735 0.6830 0.000 0.388 0.096 0.516
#> SRR1274535 4 0.7013 0.6943 0.000 0.356 0.128 0.516
#> SRR1274536 4 0.5742 0.5865 0.000 0.276 0.060 0.664
#> SRR1274537 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274539 4 0.6735 0.6830 0.000 0.388 0.096 0.516
#> SRR1274540 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.2973 0.7167 0.000 0.856 0.000 0.144
#> SRR1274546 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.4134 0.6593 0.000 0.000 0.740 0.260
#> SRR1274550 3 0.2704 0.7851 0.000 0.000 0.876 0.124
#> SRR1274551 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.4304 0.6771 0.000 0.000 0.716 0.284
#> SRR1274553 3 0.4193 0.6964 0.000 0.000 0.732 0.268
#> SRR1274554 4 0.6972 0.6954 0.000 0.356 0.124 0.520
#> SRR1274555 3 0.0188 0.8077 0.000 0.000 0.996 0.004
#> SRR1274556 3 0.0336 0.8063 0.000 0.000 0.992 0.008
#> SRR1274557 3 0.0188 0.8077 0.000 0.000 0.996 0.004
#> SRR1274558 2 0.5442 -0.0719 0.028 0.636 0.000 0.336
#> SRR1274559 1 0.5374 0.6384 0.688 0.020 0.012 0.280
#> SRR1274560 1 0.4485 0.6891 0.740 0.000 0.012 0.248
#> SRR1274561 1 0.7191 0.4166 0.516 0.000 0.328 0.156
#> SRR1274562 1 0.4761 0.4404 0.628 0.000 0.372 0.000
#> SRR1274563 3 0.0336 0.8028 0.008 0.000 0.992 0.000
#> SRR1274564 4 0.4643 0.5307 0.000 0.344 0.000 0.656
#> SRR1274565 4 0.4713 0.5128 0.000 0.360 0.000 0.640
#> SRR1274566 4 0.4643 0.5307 0.000 0.344 0.000 0.656
#> SRR1274567 4 0.5890 0.5861 0.000 0.268 0.072 0.660
#> SRR1274568 4 0.6118 0.5873 0.028 0.412 0.012 0.548
#> SRR1274569 1 0.1474 0.8191 0.948 0.000 0.000 0.052
#> SRR1274570 1 0.1302 0.8206 0.956 0.000 0.000 0.044
#> SRR1274571 1 0.1302 0.8206 0.956 0.000 0.000 0.044
#> SRR1274572 1 0.1004 0.8176 0.972 0.000 0.004 0.024
#> SRR1274573 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8895 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.4222 0.5276 0.000 0.728 0.000 0.272
#> SRR1274576 2 0.0592 0.8677 0.000 0.984 0.000 0.016
#> SRR1274577 4 0.6766 0.6889 0.000 0.380 0.100 0.520
#> SRR1274578 1 0.1661 0.8128 0.944 0.000 0.004 0.052
#> SRR1274579 1 0.1661 0.8128 0.944 0.000 0.004 0.052
#> SRR1274580 1 0.1661 0.8128 0.944 0.000 0.004 0.052
#> SRR1274581 3 0.7449 0.2171 0.332 0.000 0.480 0.188
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.4532 0.5317 0.000 0.672 0.004 0.304 0.020
#> SRR1274530 2 0.4532 0.5317 0.000 0.672 0.004 0.304 0.020
#> SRR1274531 4 0.6005 -0.3638 0.000 0.036 0.044 0.508 0.412
#> SRR1274532 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.6874 0.7546 0.000 0.136 0.052 0.264 0.548
#> SRR1274534 5 0.6853 0.7530 0.000 0.148 0.044 0.264 0.544
#> SRR1274535 5 0.6839 0.7525 0.000 0.132 0.052 0.264 0.552
#> SRR1274536 4 0.2756 0.7627 0.000 0.096 0.012 0.880 0.012
#> SRR1274537 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.6853 0.7530 0.000 0.148 0.044 0.264 0.544
#> SRR1274540 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.2144 0.8312 0.000 0.912 0.000 0.068 0.020
#> SRR1274546 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0404 0.8933 0.000 0.988 0.000 0.000 0.012
#> SRR1274548 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 4 0.4354 0.2917 0.000 0.000 0.368 0.624 0.008
#> SRR1274550 3 0.3884 0.5613 0.000 0.000 0.708 0.004 0.288
#> SRR1274551 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 5 0.5883 0.2378 0.000 0.000 0.388 0.104 0.508
#> SRR1274553 5 0.5815 0.2160 0.000 0.000 0.396 0.096 0.508
#> SRR1274554 5 0.7020 0.7379 0.000 0.140 0.052 0.292 0.516
#> SRR1274555 3 0.0290 0.9078 0.000 0.000 0.992 0.000 0.008
#> SRR1274556 3 0.0566 0.9044 0.000 0.000 0.984 0.004 0.012
#> SRR1274557 3 0.0404 0.9075 0.000 0.000 0.988 0.000 0.012
#> SRR1274558 2 0.6538 -0.2097 0.008 0.444 0.000 0.152 0.396
#> SRR1274559 1 0.6505 0.4807 0.440 0.000 0.004 0.164 0.392
#> SRR1274560 1 0.6451 0.4941 0.452 0.000 0.004 0.156 0.388
#> SRR1274561 1 0.8233 0.3686 0.332 0.000 0.296 0.116 0.256
#> SRR1274562 1 0.6029 0.2311 0.456 0.000 0.448 0.008 0.088
#> SRR1274563 3 0.0324 0.9040 0.000 0.000 0.992 0.004 0.004
#> SRR1274564 4 0.2329 0.7647 0.000 0.124 0.000 0.876 0.000
#> SRR1274565 4 0.2864 0.7479 0.000 0.136 0.000 0.852 0.012
#> SRR1274566 4 0.2329 0.7647 0.000 0.124 0.000 0.876 0.000
#> SRR1274567 4 0.2756 0.7627 0.000 0.096 0.012 0.880 0.012
#> SRR1274568 5 0.5895 0.5321 0.008 0.148 0.000 0.220 0.624
#> SRR1274569 1 0.4178 0.6775 0.748 0.000 0.004 0.028 0.220
#> SRR1274570 1 0.3951 0.6826 0.776 0.000 0.004 0.028 0.192
#> SRR1274571 1 0.3951 0.6826 0.776 0.000 0.004 0.028 0.192
#> SRR1274572 1 0.0000 0.6694 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9011 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.4532 0.5317 0.000 0.672 0.004 0.304 0.020
#> SRR1274576 2 0.1608 0.8384 0.000 0.928 0.000 0.000 0.072
#> SRR1274577 5 0.6871 0.7285 0.000 0.144 0.040 0.292 0.524
#> SRR1274578 1 0.2142 0.6547 0.920 0.000 0.004 0.028 0.048
#> SRR1274579 1 0.2142 0.6547 0.920 0.000 0.004 0.028 0.048
#> SRR1274580 1 0.2142 0.6547 0.920 0.000 0.004 0.028 0.048
#> SRR1274581 1 0.8412 0.0807 0.352 0.000 0.188 0.260 0.200
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.5798 0.0363 0.000 0.456 0.008 0.436 0.020 0.080
#> SRR1274530 2 0.5798 0.0363 0.000 0.456 0.008 0.436 0.020 0.080
#> SRR1274531 5 0.3198 0.5271 0.000 0.000 0.000 0.260 0.740 0.000
#> SRR1274532 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.1007 0.8090 0.000 0.044 0.000 0.000 0.956 0.000
#> SRR1274534 5 0.1141 0.8091 0.000 0.052 0.000 0.000 0.948 0.000
#> SRR1274535 5 0.1007 0.8090 0.000 0.044 0.000 0.000 0.956 0.000
#> SRR1274536 4 0.3235 0.9220 0.000 0.052 0.000 0.820 0.128 0.000
#> SRR1274537 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.1141 0.8091 0.000 0.052 0.000 0.000 0.948 0.000
#> SRR1274540 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.3137 0.7627 0.000 0.860 0.008 0.036 0.020 0.076
#> SRR1274546 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0603 0.8496 0.000 0.980 0.000 0.000 0.016 0.004
#> SRR1274548 2 0.0146 0.8601 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274549 4 0.5247 0.5635 0.000 0.000 0.260 0.620 0.108 0.012
#> SRR1274550 3 0.6256 0.1070 0.000 0.000 0.480 0.064 0.360 0.096
#> SRR1274551 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.5937 0.4870 0.000 0.000 0.136 0.096 0.628 0.140
#> SRR1274553 5 0.5937 0.4870 0.000 0.000 0.136 0.096 0.628 0.140
#> SRR1274554 5 0.1700 0.8056 0.000 0.048 0.000 0.024 0.928 0.000
#> SRR1274555 3 0.0632 0.7626 0.000 0.000 0.976 0.000 0.024 0.000
#> SRR1274556 3 0.1149 0.7568 0.000 0.000 0.960 0.008 0.024 0.008
#> SRR1274557 3 0.0632 0.7626 0.000 0.000 0.976 0.000 0.024 0.000
#> SRR1274558 2 0.7098 -0.1851 0.332 0.376 0.004 0.068 0.220 0.000
#> SRR1274559 1 0.5042 0.5409 0.700 0.000 0.008 0.108 0.164 0.020
#> SRR1274560 1 0.5000 0.5435 0.704 0.000 0.008 0.104 0.164 0.020
#> SRR1274561 1 0.5374 0.4742 0.656 0.000 0.188 0.132 0.008 0.016
#> SRR1274562 3 0.5462 -0.0514 0.404 0.000 0.492 0.008 0.000 0.096
#> SRR1274563 3 0.0891 0.7604 0.000 0.000 0.968 0.008 0.024 0.000
#> SRR1274564 4 0.3254 0.9214 0.000 0.056 0.000 0.820 0.124 0.000
#> SRR1274565 4 0.3627 0.9144 0.000 0.056 0.000 0.800 0.136 0.008
#> SRR1274566 4 0.3463 0.9222 0.000 0.056 0.000 0.816 0.120 0.008
#> SRR1274567 4 0.3444 0.9226 0.000 0.052 0.000 0.816 0.124 0.008
#> SRR1274568 5 0.5715 0.2830 0.328 0.040 0.004 0.068 0.560 0.000
#> SRR1274569 1 0.0000 0.5551 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.1556 0.5094 0.920 0.000 0.000 0.000 0.000 0.080
#> SRR1274571 1 0.1556 0.5094 0.920 0.000 0.000 0.000 0.000 0.080
#> SRR1274572 1 0.3804 -0.4961 0.576 0.000 0.000 0.000 0.000 0.424
#> SRR1274573 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8628 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.5798 0.0363 0.000 0.456 0.008 0.436 0.020 0.080
#> SRR1274576 2 0.2340 0.7350 0.000 0.852 0.000 0.000 0.148 0.000
#> SRR1274577 5 0.1765 0.8047 0.000 0.052 0.000 0.024 0.924 0.000
#> SRR1274578 6 0.3804 0.6906 0.424 0.000 0.000 0.000 0.000 0.576
#> SRR1274579 6 0.3804 0.6906 0.424 0.000 0.000 0.000 0.000 0.576
#> SRR1274580 6 0.3804 0.6906 0.424 0.000 0.000 0.000 0.000 0.576
#> SRR1274581 6 0.4658 0.2658 0.012 0.000 0.084 0.140 0.020 0.744
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "skmeans"]
# you can also extract it by
# res = res_list["SD:skmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'skmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.923 0.937 0.973 0.5016 0.502 0.502
#> 3 3 0.894 0.924 0.965 0.3182 0.718 0.496
#> 4 4 0.853 0.868 0.936 0.1223 0.866 0.627
#> 5 5 0.835 0.832 0.913 0.0627 0.920 0.705
#> 6 6 0.814 0.808 0.888 0.0308 0.962 0.824
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.964 0.000 1.000
#> SRR1274529 2 0.000 0.964 0.000 1.000
#> SRR1274530 2 0.000 0.964 0.000 1.000
#> SRR1274531 2 0.706 0.766 0.192 0.808
#> SRR1274532 2 0.000 0.964 0.000 1.000
#> SRR1274533 2 0.000 0.964 0.000 1.000
#> SRR1274534 2 0.000 0.964 0.000 1.000
#> SRR1274535 2 0.996 0.168 0.464 0.536
#> SRR1274536 2 0.615 0.816 0.152 0.848
#> SRR1274537 2 0.000 0.964 0.000 1.000
#> SRR1274538 2 0.000 0.964 0.000 1.000
#> SRR1274539 2 0.000 0.964 0.000 1.000
#> SRR1274540 2 0.000 0.964 0.000 1.000
#> SRR1274541 2 0.000 0.964 0.000 1.000
#> SRR1274542 2 0.000 0.964 0.000 1.000
#> SRR1274543 2 0.000 0.964 0.000 1.000
#> SRR1274544 2 0.000 0.964 0.000 1.000
#> SRR1274545 2 0.000 0.964 0.000 1.000
#> SRR1274546 2 0.000 0.964 0.000 1.000
#> SRR1274547 2 0.000 0.964 0.000 1.000
#> SRR1274548 2 0.000 0.964 0.000 1.000
#> SRR1274549 1 0.000 0.981 1.000 0.000
#> SRR1274550 1 0.000 0.981 1.000 0.000
#> SRR1274551 2 0.000 0.964 0.000 1.000
#> SRR1274552 1 0.000 0.981 1.000 0.000
#> SRR1274553 1 0.000 0.981 1.000 0.000
#> SRR1274554 2 0.224 0.935 0.036 0.964
#> SRR1274555 1 0.000 0.981 1.000 0.000
#> SRR1274556 1 0.000 0.981 1.000 0.000
#> SRR1274557 1 0.000 0.981 1.000 0.000
#> SRR1274558 2 0.730 0.736 0.204 0.796
#> SRR1274559 1 0.000 0.981 1.000 0.000
#> SRR1274560 1 0.000 0.981 1.000 0.000
#> SRR1274561 1 0.000 0.981 1.000 0.000
#> SRR1274562 1 0.000 0.981 1.000 0.000
#> SRR1274563 1 0.000 0.981 1.000 0.000
#> SRR1274564 2 0.000 0.964 0.000 1.000
#> SRR1274565 2 0.000 0.964 0.000 1.000
#> SRR1274566 2 0.000 0.964 0.000 1.000
#> SRR1274567 1 0.788 0.674 0.764 0.236
#> SRR1274568 1 0.388 0.911 0.924 0.076
#> SRR1274569 1 0.000 0.981 1.000 0.000
#> SRR1274570 1 0.000 0.981 1.000 0.000
#> SRR1274571 1 0.000 0.981 1.000 0.000
#> SRR1274572 1 0.000 0.981 1.000 0.000
#> SRR1274573 2 0.000 0.964 0.000 1.000
#> SRR1274574 2 0.000 0.964 0.000 1.000
#> SRR1274575 2 0.000 0.964 0.000 1.000
#> SRR1274576 2 0.000 0.964 0.000 1.000
#> SRR1274577 1 0.388 0.911 0.924 0.076
#> SRR1274578 1 0.000 0.981 1.000 0.000
#> SRR1274579 1 0.000 0.981 1.000 0.000
#> SRR1274580 1 0.000 0.981 1.000 0.000
#> SRR1274581 1 0.000 0.981 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274529 2 0.0747 0.979 0.000 0.984 0.016
#> SRR1274530 2 0.0747 0.979 0.000 0.984 0.016
#> SRR1274531 3 0.0000 0.932 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274533 3 0.0747 0.931 0.000 0.016 0.984
#> SRR1274534 3 0.4887 0.731 0.000 0.228 0.772
#> SRR1274535 3 0.0747 0.931 0.000 0.016 0.984
#> SRR1274536 3 0.0000 0.932 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274539 3 0.4887 0.731 0.000 0.228 0.772
#> SRR1274540 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274545 2 0.0747 0.979 0.000 0.984 0.016
#> SRR1274546 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.932 0.000 0.000 1.000
#> SRR1274550 3 0.0747 0.934 0.016 0.000 0.984
#> SRR1274551 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274552 3 0.0892 0.933 0.020 0.000 0.980
#> SRR1274553 3 0.0892 0.933 0.020 0.000 0.980
#> SRR1274554 3 0.0747 0.931 0.000 0.016 0.984
#> SRR1274555 3 0.0747 0.934 0.016 0.000 0.984
#> SRR1274556 3 0.0747 0.934 0.016 0.000 0.984
#> SRR1274557 3 0.0747 0.934 0.016 0.000 0.984
#> SRR1274558 1 0.6045 0.386 0.620 0.380 0.000
#> SRR1274559 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274560 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274561 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274562 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274563 3 0.1163 0.928 0.028 0.000 0.972
#> SRR1274564 2 0.3192 0.887 0.000 0.888 0.112
#> SRR1274565 2 0.0747 0.979 0.000 0.984 0.016
#> SRR1274566 2 0.3038 0.896 0.000 0.896 0.104
#> SRR1274567 3 0.0237 0.932 0.004 0.000 0.996
#> SRR1274568 1 0.0747 0.934 0.984 0.016 0.000
#> SRR1274569 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274575 2 0.0747 0.979 0.000 0.984 0.016
#> SRR1274576 2 0.0000 0.986 0.000 1.000 0.000
#> SRR1274577 3 0.6629 0.436 0.360 0.016 0.624
#> SRR1274578 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.949 1.000 0.000 0.000
#> SRR1274581 1 0.4291 0.750 0.820 0.000 0.180
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274529 4 0.4898 0.476 0.000 0.416 0.000 0.584
#> SRR1274530 4 0.4898 0.476 0.000 0.416 0.000 0.584
#> SRR1274531 3 0.3688 0.812 0.000 0.000 0.792 0.208
#> SRR1274532 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.0000 0.917 0.000 0.000 1.000 0.000
#> SRR1274534 3 0.2149 0.855 0.000 0.088 0.912 0.000
#> SRR1274535 3 0.0000 0.917 0.000 0.000 1.000 0.000
#> SRR1274536 4 0.0000 0.747 0.000 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274539 3 0.2149 0.855 0.000 0.088 0.912 0.000
#> SRR1274540 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.1302 0.911 0.000 0.956 0.000 0.044
#> SRR1274546 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274549 4 0.3726 0.502 0.000 0.000 0.212 0.788
#> SRR1274550 3 0.0336 0.918 0.000 0.000 0.992 0.008
#> SRR1274551 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.0336 0.918 0.000 0.000 0.992 0.008
#> SRR1274553 3 0.0336 0.918 0.000 0.000 0.992 0.008
#> SRR1274554 3 0.0469 0.915 0.000 0.000 0.988 0.012
#> SRR1274555 3 0.3160 0.880 0.020 0.000 0.872 0.108
#> SRR1274556 3 0.3658 0.858 0.020 0.000 0.836 0.144
#> SRR1274557 3 0.2469 0.887 0.000 0.000 0.892 0.108
#> SRR1274558 2 0.4907 0.255 0.420 0.580 0.000 0.000
#> SRR1274559 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.0592 0.960 0.984 0.000 0.000 0.016
#> SRR1274562 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274563 3 0.4608 0.826 0.096 0.000 0.800 0.104
#> SRR1274564 4 0.0336 0.751 0.000 0.008 0.000 0.992
#> SRR1274565 4 0.2654 0.743 0.000 0.108 0.004 0.888
#> SRR1274566 4 0.0376 0.750 0.000 0.004 0.004 0.992
#> SRR1274567 4 0.0000 0.747 0.000 0.000 0.000 1.000
#> SRR1274568 1 0.1767 0.927 0.944 0.012 0.044 0.000
#> SRR1274569 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.962 0.000 1.000 0.000 0.000
#> SRR1274575 4 0.4898 0.476 0.000 0.416 0.000 0.584
#> SRR1274576 2 0.0188 0.958 0.000 0.996 0.004 0.000
#> SRR1274577 3 0.1610 0.901 0.032 0.000 0.952 0.016
#> SRR1274578 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.972 1.000 0.000 0.000 0.000
#> SRR1274581 1 0.5109 0.674 0.736 0.000 0.052 0.212
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 4 0.3177 0.792 0.000 0.208 0.000 0.792 0.000
#> SRR1274530 4 0.3177 0.792 0.000 0.208 0.000 0.792 0.000
#> SRR1274531 3 0.5382 0.556 0.000 0.000 0.656 0.120 0.224
#> SRR1274532 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.1732 0.962 0.000 0.000 0.080 0.000 0.920
#> SRR1274534 5 0.1670 0.969 0.000 0.012 0.052 0.000 0.936
#> SRR1274535 5 0.1671 0.966 0.000 0.000 0.076 0.000 0.924
#> SRR1274536 4 0.0290 0.872 0.000 0.000 0.008 0.992 0.000
#> SRR1274537 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.1670 0.969 0.000 0.012 0.052 0.000 0.936
#> SRR1274540 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.2561 0.800 0.000 0.856 0.000 0.144 0.000
#> SRR1274546 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 3 0.3424 0.678 0.000 0.000 0.760 0.240 0.000
#> SRR1274550 3 0.2127 0.796 0.000 0.000 0.892 0.000 0.108
#> SRR1274551 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.2732 0.764 0.000 0.000 0.840 0.000 0.160
#> SRR1274553 3 0.2732 0.764 0.000 0.000 0.840 0.000 0.160
#> SRR1274554 5 0.1478 0.970 0.000 0.000 0.064 0.000 0.936
#> SRR1274555 3 0.0000 0.821 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0000 0.821 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.0880 0.818 0.000 0.000 0.968 0.000 0.032
#> SRR1274558 1 0.4708 0.211 0.548 0.436 0.000 0.000 0.016
#> SRR1274559 1 0.0510 0.838 0.984 0.000 0.000 0.000 0.016
#> SRR1274560 1 0.0510 0.838 0.984 0.000 0.000 0.000 0.016
#> SRR1274561 1 0.3039 0.745 0.836 0.000 0.152 0.000 0.012
#> SRR1274562 1 0.5151 0.539 0.644 0.000 0.300 0.008 0.048
#> SRR1274563 3 0.0000 0.821 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.0290 0.872 0.000 0.000 0.008 0.992 0.000
#> SRR1274565 4 0.0290 0.871 0.000 0.008 0.000 0.992 0.000
#> SRR1274566 4 0.0290 0.872 0.000 0.000 0.008 0.992 0.000
#> SRR1274567 4 0.0290 0.872 0.000 0.000 0.008 0.992 0.000
#> SRR1274568 1 0.4192 0.294 0.596 0.000 0.000 0.000 0.404
#> SRR1274569 1 0.0162 0.840 0.996 0.000 0.000 0.000 0.004
#> SRR1274570 1 0.0162 0.841 0.996 0.000 0.004 0.000 0.000
#> SRR1274571 1 0.0162 0.841 0.996 0.000 0.004 0.000 0.000
#> SRR1274572 1 0.2352 0.834 0.912 0.000 0.032 0.008 0.048
#> SRR1274573 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.970 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.3177 0.792 0.000 0.208 0.000 0.792 0.000
#> SRR1274576 2 0.3684 0.605 0.000 0.720 0.000 0.000 0.280
#> SRR1274577 5 0.0566 0.928 0.004 0.000 0.012 0.000 0.984
#> SRR1274578 1 0.2352 0.834 0.912 0.000 0.032 0.008 0.048
#> SRR1274579 1 0.2352 0.834 0.912 0.000 0.032 0.008 0.048
#> SRR1274580 1 0.2352 0.834 0.912 0.000 0.032 0.008 0.048
#> SRR1274581 3 0.5302 0.298 0.336 0.000 0.608 0.008 0.048
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.4563 0.658 0.056 0.284 0.000 0.656 0.004 0.000
#> SRR1274530 4 0.4581 0.653 0.056 0.288 0.000 0.652 0.004 0.000
#> SRR1274531 3 0.4556 0.666 0.000 0.000 0.696 0.188 0.116 0.000
#> SRR1274532 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.1552 0.919 0.036 0.000 0.020 0.000 0.940 0.004
#> SRR1274534 5 0.0363 0.936 0.000 0.012 0.000 0.000 0.988 0.000
#> SRR1274535 5 0.1498 0.916 0.028 0.000 0.032 0.000 0.940 0.000
#> SRR1274536 4 0.0146 0.782 0.000 0.000 0.004 0.996 0.000 0.000
#> SRR1274537 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.0363 0.936 0.000 0.012 0.000 0.000 0.988 0.000
#> SRR1274540 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.3079 0.804 0.056 0.844 0.000 0.096 0.004 0.000
#> SRR1274546 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.1429 0.921 0.052 0.940 0.000 0.004 0.004 0.000
#> SRR1274548 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 3 0.3482 0.595 0.000 0.000 0.684 0.316 0.000 0.000
#> SRR1274550 3 0.2917 0.772 0.016 0.000 0.840 0.000 0.136 0.008
#> SRR1274551 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 3 0.4154 0.674 0.036 0.000 0.712 0.000 0.244 0.008
#> SRR1274553 3 0.4130 0.678 0.036 0.000 0.716 0.000 0.240 0.008
#> SRR1274554 5 0.1555 0.918 0.040 0.000 0.012 0.008 0.940 0.000
#> SRR1274555 3 0.0146 0.818 0.004 0.000 0.996 0.000 0.000 0.000
#> SRR1274556 3 0.0000 0.818 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274557 3 0.0146 0.818 0.004 0.000 0.996 0.000 0.000 0.000
#> SRR1274558 1 0.2775 0.648 0.856 0.104 0.000 0.000 0.000 0.040
#> SRR1274559 1 0.2378 0.728 0.848 0.000 0.000 0.000 0.000 0.152
#> SRR1274560 1 0.2378 0.728 0.848 0.000 0.000 0.000 0.000 0.152
#> SRR1274561 1 0.5336 0.592 0.592 0.000 0.180 0.000 0.000 0.228
#> SRR1274562 6 0.3954 0.481 0.012 0.000 0.352 0.000 0.000 0.636
#> SRR1274563 3 0.0146 0.818 0.004 0.000 0.996 0.000 0.000 0.000
#> SRR1274564 4 0.0146 0.782 0.000 0.000 0.004 0.996 0.000 0.000
#> SRR1274565 4 0.1049 0.775 0.032 0.000 0.000 0.960 0.008 0.000
#> SRR1274566 4 0.0717 0.781 0.016 0.000 0.000 0.976 0.008 0.000
#> SRR1274567 4 0.0717 0.781 0.016 0.000 0.000 0.976 0.008 0.000
#> SRR1274568 1 0.2560 0.657 0.872 0.000 0.000 0.000 0.092 0.036
#> SRR1274569 1 0.3804 0.610 0.576 0.000 0.000 0.000 0.000 0.424
#> SRR1274570 1 0.3851 0.564 0.540 0.000 0.000 0.000 0.000 0.460
#> SRR1274571 1 0.3847 0.571 0.544 0.000 0.000 0.000 0.000 0.456
#> SRR1274572 6 0.0713 0.815 0.028 0.000 0.000 0.000 0.000 0.972
#> SRR1274573 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 4 0.4581 0.653 0.056 0.288 0.000 0.652 0.004 0.000
#> SRR1274576 2 0.3352 0.738 0.032 0.792 0.000 0.000 0.176 0.000
#> SRR1274577 5 0.2928 0.856 0.056 0.000 0.000 0.004 0.856 0.084
#> SRR1274578 6 0.0260 0.832 0.008 0.000 0.000 0.000 0.000 0.992
#> SRR1274579 6 0.0260 0.832 0.008 0.000 0.000 0.000 0.000 0.992
#> SRR1274580 6 0.0260 0.832 0.008 0.000 0.000 0.000 0.000 0.992
#> SRR1274581 6 0.2809 0.701 0.004 0.000 0.168 0.004 0.000 0.824
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "pam"]
# you can also extract it by
# res = res_list["SD:pam"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'pam' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 6.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.820 0.892 0.953 0.4152 0.609 0.609
#> 3 3 0.911 0.876 0.936 0.2619 0.844 0.748
#> 4 4 0.831 0.846 0.921 0.1367 0.957 0.910
#> 5 5 1.000 0.948 0.980 0.1743 0.822 0.612
#> 6 6 1.000 0.974 0.989 0.0487 0.957 0.859
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 6
#> attr(,"optional")
#> [1] 3 5
There is also optional best \(k\) = 3 5 that is worth to check.
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.938 0.000 1.000
#> SRR1274529 2 0.000 0.938 0.000 1.000
#> SRR1274530 2 0.000 0.938 0.000 1.000
#> SRR1274531 2 0.416 0.871 0.084 0.916
#> SRR1274532 2 0.000 0.938 0.000 1.000
#> SRR1274533 2 0.000 0.938 0.000 1.000
#> SRR1274534 2 0.000 0.938 0.000 1.000
#> SRR1274535 2 0.000 0.938 0.000 1.000
#> SRR1274536 2 0.861 0.647 0.284 0.716
#> SRR1274537 2 0.000 0.938 0.000 1.000
#> SRR1274538 2 0.000 0.938 0.000 1.000
#> SRR1274539 2 0.000 0.938 0.000 1.000
#> SRR1274540 2 0.000 0.938 0.000 1.000
#> SRR1274541 2 0.000 0.938 0.000 1.000
#> SRR1274542 2 0.000 0.938 0.000 1.000
#> SRR1274543 2 0.000 0.938 0.000 1.000
#> SRR1274544 2 0.000 0.938 0.000 1.000
#> SRR1274545 2 0.000 0.938 0.000 1.000
#> SRR1274546 2 0.000 0.938 0.000 1.000
#> SRR1274547 2 0.000 0.938 0.000 1.000
#> SRR1274548 2 0.000 0.938 0.000 1.000
#> SRR1274549 2 0.866 0.641 0.288 0.712
#> SRR1274550 2 0.866 0.641 0.288 0.712
#> SRR1274551 2 0.000 0.938 0.000 1.000
#> SRR1274552 2 0.861 0.647 0.284 0.716
#> SRR1274553 2 0.876 0.628 0.296 0.704
#> SRR1274554 2 0.000 0.938 0.000 1.000
#> SRR1274555 1 0.000 0.971 1.000 0.000
#> SRR1274556 2 0.980 0.369 0.416 0.584
#> SRR1274557 1 0.921 0.410 0.664 0.336
#> SRR1274558 2 0.000 0.938 0.000 1.000
#> SRR1274559 1 0.000 0.971 1.000 0.000
#> SRR1274560 1 0.000 0.971 1.000 0.000
#> SRR1274561 1 0.000 0.971 1.000 0.000
#> SRR1274562 1 0.000 0.971 1.000 0.000
#> SRR1274563 1 0.000 0.971 1.000 0.000
#> SRR1274564 2 0.000 0.938 0.000 1.000
#> SRR1274565 2 0.000 0.938 0.000 1.000
#> SRR1274566 2 0.000 0.938 0.000 1.000
#> SRR1274567 2 0.866 0.641 0.288 0.712
#> SRR1274568 2 0.000 0.938 0.000 1.000
#> SRR1274569 1 0.000 0.971 1.000 0.000
#> SRR1274570 1 0.000 0.971 1.000 0.000
#> SRR1274571 1 0.000 0.971 1.000 0.000
#> SRR1274572 1 0.000 0.971 1.000 0.000
#> SRR1274573 2 0.000 0.938 0.000 1.000
#> SRR1274574 2 0.000 0.938 0.000 1.000
#> SRR1274575 2 0.000 0.938 0.000 1.000
#> SRR1274576 2 0.000 0.938 0.000 1.000
#> SRR1274577 2 0.000 0.938 0.000 1.000
#> SRR1274578 1 0.000 0.971 1.000 0.000
#> SRR1274579 1 0.000 0.971 1.000 0.000
#> SRR1274580 1 0.000 0.971 1.000 0.000
#> SRR1274581 2 0.000 0.938 0.000 1.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274529 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274530 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274531 2 0.280 0.871 0.000 0.908 0.092
#> SRR1274532 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274533 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274534 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274535 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274536 2 0.536 0.570 0.000 0.724 0.276
#> SRR1274537 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274538 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274539 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274540 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274541 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274542 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274543 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274544 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274545 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274546 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274547 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274548 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274549 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274550 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274551 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274552 3 0.629 0.110 0.000 0.464 0.536
#> SRR1274553 3 0.271 0.732 0.000 0.088 0.912
#> SRR1274554 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274555 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274556 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274557 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274558 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274559 1 0.536 0.824 0.724 0.000 0.276
#> SRR1274560 1 0.536 0.824 0.724 0.000 0.276
#> SRR1274561 1 0.536 0.824 0.724 0.000 0.276
#> SRR1274562 1 0.581 0.756 0.664 0.000 0.336
#> SRR1274563 3 0.000 0.840 0.000 0.000 1.000
#> SRR1274564 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274565 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274566 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274567 2 0.536 0.570 0.000 0.724 0.276
#> SRR1274568 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274569 1 0.536 0.824 0.724 0.000 0.276
#> SRR1274570 1 0.518 0.823 0.744 0.000 0.256
#> SRR1274571 1 0.536 0.824 0.724 0.000 0.276
#> SRR1274572 1 0.000 0.750 1.000 0.000 0.000
#> SRR1274573 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274574 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274575 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274576 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274577 2 0.000 0.969 0.000 1.000 0.000
#> SRR1274578 1 0.000 0.750 1.000 0.000 0.000
#> SRR1274579 1 0.000 0.750 1.000 0.000 0.000
#> SRR1274580 1 0.000 0.750 1.000 0.000 0.000
#> SRR1274581 2 0.536 0.649 0.276 0.724 0.000
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274531 2 0.7336 0.360 0.284 0.520 0.196 0.000
#> SRR1274532 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274533 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274534 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274535 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274536 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274537 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274539 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274540 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.4304 0.547 0.284 0.000 0.716 0.000
#> SRR1274550 3 0.0000 0.814 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.5821 0.272 0.032 0.432 0.536 0.000
#> SRR1274553 3 0.2149 0.741 0.000 0.088 0.912 0.000
#> SRR1274554 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274555 3 0.0000 0.814 0.000 0.000 1.000 0.000
#> SRR1274556 3 0.0000 0.814 0.000 0.000 1.000 0.000
#> SRR1274557 3 0.0000 0.814 0.000 0.000 1.000 0.000
#> SRR1274558 2 0.0336 0.914 0.008 0.992 0.000 0.000
#> SRR1274559 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274560 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274561 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274562 1 0.5496 0.903 0.704 0.000 0.064 0.232
#> SRR1274563 3 0.0000 0.814 0.000 0.000 1.000 0.000
#> SRR1274564 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274565 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274566 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274567 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274568 2 0.0336 0.914 0.008 0.992 0.000 0.000
#> SRR1274569 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274570 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274571 1 0.4304 0.985 0.716 0.000 0.000 0.284
#> SRR1274572 1 0.4356 0.977 0.708 0.000 0.000 0.292
#> SRR1274573 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274576 2 0.0000 0.920 0.000 1.000 0.000 0.000
#> SRR1274577 2 0.4304 0.709 0.284 0.716 0.000 0.000
#> SRR1274578 4 0.0000 0.825 0.000 0.000 0.000 1.000
#> SRR1274579 4 0.0000 0.825 0.000 0.000 0.000 1.000
#> SRR1274580 4 0.0000 0.825 0.000 0.000 0.000 1.000
#> SRR1274581 4 0.4304 0.568 0.284 0.000 0.000 0.716
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274530 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274531 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274532 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274534 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274535 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274536 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274537 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274540 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274550 3 0.000 0.968 0.000 0.000 1.000 0.000 0.000
#> SRR1274551 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 2 0.499 0.335 0.000 0.600 0.360 0.040 0.000
#> SRR1274553 3 0.185 0.835 0.000 0.088 0.912 0.000 0.000
#> SRR1274554 4 0.179 0.855 0.000 0.084 0.000 0.916 0.000
#> SRR1274555 3 0.000 0.968 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.000 0.968 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.000 0.968 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 2 0.148 0.919 0.064 0.936 0.000 0.000 0.000
#> SRR1274559 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274562 1 0.196 0.892 0.904 0.000 0.096 0.000 0.000
#> SRR1274563 3 0.000 0.968 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274565 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274566 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274567 4 0.000 0.973 0.000 0.000 0.000 1.000 0.000
#> SRR1274568 2 0.148 0.919 0.064 0.936 0.000 0.000 0.000
#> SRR1274569 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.000 0.975 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.148 0.927 0.936 0.000 0.000 0.000 0.064
#> SRR1274573 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274576 2 0.000 0.979 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 4 0.104 0.927 0.000 0.040 0.000 0.960 0.000
#> SRR1274578 5 0.000 0.930 0.000 0.000 0.000 0.000 1.000
#> SRR1274579 5 0.000 0.930 0.000 0.000 0.000 0.000 1.000
#> SRR1274580 5 0.000 0.930 0.000 0.000 0.000 0.000 1.000
#> SRR1274581 5 0.293 0.776 0.000 0.000 0.000 0.180 0.820
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274530 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274531 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274532 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 2 0.0713 0.969 0.000 0.972 0.000 0.000 0.028 0.000
#> SRR1274534 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274535 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274536 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274537 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274540 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274550 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274553 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274554 4 0.1610 0.859 0.000 0.084 0.000 0.916 0.000 0.000
#> SRR1274555 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274557 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274558 2 0.1327 0.934 0.064 0.936 0.000 0.000 0.000 0.000
#> SRR1274559 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274562 1 0.1765 0.897 0.904 0.000 0.096 0.000 0.000 0.000
#> SRR1274563 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274565 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274566 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274567 4 0.0000 0.974 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274568 2 0.1327 0.934 0.064 0.936 0.000 0.000 0.000 0.000
#> SRR1274569 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.976 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.1327 0.928 0.936 0.000 0.000 0.000 0.000 0.064
#> SRR1274573 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274576 2 0.0000 0.993 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 4 0.0937 0.929 0.000 0.040 0.000 0.960 0.000 0.000
#> SRR1274578 6 0.0000 0.935 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274579 6 0.0000 0.935 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274580 6 0.0000 0.935 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274581 6 0.3235 0.791 0.000 0.000 0.000 0.128 0.052 0.820
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "mclust"]
# you can also extract it by
# res = res_list["SD:mclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'mclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.759 0.950 0.965 0.4504 0.525 0.525
#> 3 3 0.430 0.467 0.719 0.3199 0.709 0.490
#> 4 4 0.473 0.706 0.767 0.1476 0.778 0.504
#> 5 5 0.608 0.402 0.667 0.1466 0.845 0.574
#> 6 6 0.712 0.607 0.745 0.0255 0.847 0.454
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.910 0.000 1.000
#> SRR1274529 2 0.5629 0.891 0.132 0.868
#> SRR1274530 2 0.5629 0.891 0.132 0.868
#> SRR1274531 1 0.0938 0.992 0.988 0.012
#> SRR1274532 2 0.0000 0.910 0.000 1.000
#> SRR1274533 1 0.0938 0.992 0.988 0.012
#> SRR1274534 1 0.0672 0.994 0.992 0.008
#> SRR1274535 1 0.0938 0.992 0.988 0.012
#> SRR1274536 1 0.0938 0.992 0.988 0.012
#> SRR1274537 2 0.7745 0.793 0.228 0.772
#> SRR1274538 2 0.6438 0.864 0.164 0.836
#> SRR1274539 1 0.0672 0.994 0.992 0.008
#> SRR1274540 2 0.0000 0.910 0.000 1.000
#> SRR1274541 2 0.0000 0.910 0.000 1.000
#> SRR1274542 2 0.0000 0.910 0.000 1.000
#> SRR1274543 2 0.9323 0.576 0.348 0.652
#> SRR1274544 2 0.0000 0.910 0.000 1.000
#> SRR1274545 2 0.5629 0.891 0.132 0.868
#> SRR1274546 2 0.0000 0.910 0.000 1.000
#> SRR1274547 2 0.4690 0.898 0.100 0.900
#> SRR1274548 2 0.6973 0.841 0.188 0.812
#> SRR1274549 1 0.0376 0.996 0.996 0.004
#> SRR1274550 1 0.0376 0.996 0.996 0.004
#> SRR1274551 2 0.0000 0.910 0.000 1.000
#> SRR1274552 1 0.0376 0.996 0.996 0.004
#> SRR1274553 1 0.0376 0.996 0.996 0.004
#> SRR1274554 1 0.0672 0.994 0.992 0.008
#> SRR1274555 1 0.0376 0.996 0.996 0.004
#> SRR1274556 1 0.0376 0.996 0.996 0.004
#> SRR1274557 1 0.0376 0.996 0.996 0.004
#> SRR1274558 1 0.0000 0.995 1.000 0.000
#> SRR1274559 1 0.0000 0.995 1.000 0.000
#> SRR1274560 1 0.0000 0.995 1.000 0.000
#> SRR1274561 1 0.0000 0.995 1.000 0.000
#> SRR1274562 1 0.0000 0.995 1.000 0.000
#> SRR1274563 1 0.0376 0.996 0.996 0.004
#> SRR1274564 1 0.0938 0.992 0.988 0.012
#> SRR1274565 1 0.1184 0.988 0.984 0.016
#> SRR1274566 1 0.0938 0.992 0.988 0.012
#> SRR1274567 1 0.0672 0.994 0.992 0.008
#> SRR1274568 1 0.0000 0.995 1.000 0.000
#> SRR1274569 1 0.0000 0.995 1.000 0.000
#> SRR1274570 1 0.0000 0.995 1.000 0.000
#> SRR1274571 1 0.0000 0.995 1.000 0.000
#> SRR1274572 1 0.0000 0.995 1.000 0.000
#> SRR1274573 2 0.0000 0.910 0.000 1.000
#> SRR1274574 2 0.0672 0.910 0.008 0.992
#> SRR1274575 2 0.5629 0.891 0.132 0.868
#> SRR1274576 2 0.5946 0.883 0.144 0.856
#> SRR1274577 1 0.0376 0.996 0.996 0.004
#> SRR1274578 1 0.0000 0.995 1.000 0.000
#> SRR1274579 1 0.0000 0.995 1.000 0.000
#> SRR1274580 1 0.0000 0.995 1.000 0.000
#> SRR1274581 1 0.0000 0.995 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.1411 0.8498 0.000 0.964 0.036
#> SRR1274529 3 0.5467 0.4535 0.032 0.176 0.792
#> SRR1274530 3 0.5521 0.4521 0.032 0.180 0.788
#> SRR1274531 3 0.7346 0.3174 0.368 0.040 0.592
#> SRR1274532 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274533 3 0.7291 0.3261 0.356 0.040 0.604
#> SRR1274534 3 0.6228 0.2626 0.372 0.004 0.624
#> SRR1274535 3 0.7190 0.3216 0.356 0.036 0.608
#> SRR1274536 3 0.3530 0.5019 0.068 0.032 0.900
#> SRR1274537 2 0.5956 0.6396 0.004 0.672 0.324
#> SRR1274538 2 0.5760 0.6351 0.000 0.672 0.328
#> SRR1274539 3 0.6379 0.2726 0.368 0.008 0.624
#> SRR1274540 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274543 2 0.6600 0.4898 0.012 0.604 0.384
#> SRR1274544 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274545 3 0.6099 0.3856 0.032 0.228 0.740
#> SRR1274546 2 0.0237 0.8561 0.000 0.996 0.004
#> SRR1274547 2 0.5803 0.6956 0.016 0.736 0.248
#> SRR1274548 2 0.4887 0.7570 0.000 0.772 0.228
#> SRR1274549 1 0.7072 0.0839 0.504 0.020 0.476
#> SRR1274550 3 0.7186 -0.0919 0.476 0.024 0.500
#> SRR1274551 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274552 3 0.6442 0.0521 0.432 0.004 0.564
#> SRR1274553 3 0.6520 -0.1501 0.488 0.004 0.508
#> SRR1274554 3 0.7291 0.3261 0.356 0.040 0.604
#> SRR1274555 1 0.6308 0.1441 0.508 0.000 0.492
#> SRR1274556 1 0.7069 0.0970 0.508 0.020 0.472
#> SRR1274557 3 0.7074 -0.1101 0.480 0.020 0.500
#> SRR1274558 1 0.6235 0.2965 0.564 0.000 0.436
#> SRR1274559 1 0.5882 0.4267 0.652 0.000 0.348
#> SRR1274560 1 0.5397 0.4764 0.720 0.000 0.280
#> SRR1274561 1 0.6235 0.3310 0.564 0.000 0.436
#> SRR1274562 1 0.6235 0.3310 0.564 0.000 0.436
#> SRR1274563 1 0.6235 0.3310 0.564 0.000 0.436
#> SRR1274564 3 0.3649 0.5035 0.068 0.036 0.896
#> SRR1274565 3 0.3967 0.5042 0.072 0.044 0.884
#> SRR1274566 3 0.3856 0.5046 0.072 0.040 0.888
#> SRR1274567 3 0.4418 0.4742 0.132 0.020 0.848
#> SRR1274568 1 0.5882 0.4344 0.652 0.000 0.348
#> SRR1274569 1 0.0237 0.5399 0.996 0.000 0.004
#> SRR1274570 1 0.0000 0.5379 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.5379 1.000 0.000 0.000
#> SRR1274572 1 0.0424 0.5417 0.992 0.000 0.008
#> SRR1274573 2 0.0000 0.8563 0.000 1.000 0.000
#> SRR1274574 2 0.2796 0.8345 0.000 0.908 0.092
#> SRR1274575 3 0.5521 0.4521 0.032 0.180 0.788
#> SRR1274576 2 0.5291 0.7115 0.000 0.732 0.268
#> SRR1274577 3 0.6280 0.0265 0.460 0.000 0.540
#> SRR1274578 1 0.0424 0.5417 0.992 0.000 0.008
#> SRR1274579 1 0.0424 0.5417 0.992 0.000 0.008
#> SRR1274580 1 0.0424 0.5417 0.992 0.000 0.008
#> SRR1274581 1 0.6244 0.2499 0.560 0.000 0.440
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.2256 0.744 0.000 0.924 0.056 0.020
#> SRR1274529 4 0.6289 0.949 0.000 0.236 0.116 0.648
#> SRR1274530 4 0.6267 0.953 0.000 0.240 0.112 0.648
#> SRR1274531 3 0.4625 0.653 0.008 0.092 0.812 0.088
#> SRR1274532 2 0.0000 0.758 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.2565 0.657 0.000 0.032 0.912 0.056
#> SRR1274534 3 0.2313 0.667 0.000 0.032 0.924 0.044
#> SRR1274535 3 0.2443 0.658 0.000 0.024 0.916 0.060
#> SRR1274536 3 0.6705 0.494 0.012 0.080 0.592 0.316
#> SRR1274537 2 0.6041 0.490 0.000 0.608 0.332 0.060
#> SRR1274538 2 0.5759 0.546 0.000 0.668 0.268 0.064
#> SRR1274539 3 0.2111 0.667 0.000 0.024 0.932 0.044
#> SRR1274540 2 0.0000 0.758 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0469 0.757 0.000 0.988 0.012 0.000
#> SRR1274542 2 0.0188 0.759 0.000 0.996 0.004 0.000
#> SRR1274543 2 0.5973 0.480 0.000 0.612 0.332 0.056
#> SRR1274544 2 0.0188 0.759 0.000 0.996 0.004 0.000
#> SRR1274545 4 0.6801 0.856 0.000 0.308 0.124 0.568
#> SRR1274546 2 0.0817 0.759 0.000 0.976 0.024 0.000
#> SRR1274547 2 0.5440 0.522 0.000 0.736 0.104 0.160
#> SRR1274548 2 0.5905 0.527 0.000 0.636 0.304 0.060
#> SRR1274549 3 0.8105 0.650 0.172 0.064 0.564 0.200
#> SRR1274550 3 0.5172 0.715 0.160 0.020 0.772 0.048
#> SRR1274551 2 0.0000 0.758 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.2686 0.708 0.032 0.012 0.916 0.040
#> SRR1274553 3 0.4033 0.716 0.116 0.004 0.836 0.044
#> SRR1274554 3 0.2546 0.657 0.000 0.028 0.912 0.060
#> SRR1274555 3 0.6511 0.693 0.172 0.000 0.640 0.188
#> SRR1274556 3 0.6750 0.692 0.168 0.004 0.628 0.200
#> SRR1274557 3 0.6827 0.707 0.160 0.048 0.680 0.112
#> SRR1274558 3 0.5794 0.653 0.104 0.012 0.732 0.152
#> SRR1274559 3 0.6910 0.578 0.252 0.000 0.584 0.164
#> SRR1274560 3 0.7049 0.532 0.300 0.000 0.548 0.152
#> SRR1274561 3 0.6860 0.657 0.244 0.000 0.592 0.164
#> SRR1274562 3 0.7128 0.636 0.260 0.000 0.556 0.184
#> SRR1274563 3 0.6110 0.700 0.176 0.000 0.680 0.144
#> SRR1274564 3 0.6607 0.480 0.008 0.084 0.600 0.308
#> SRR1274565 3 0.6636 0.483 0.008 0.108 0.628 0.256
#> SRR1274566 3 0.6732 0.466 0.008 0.096 0.596 0.300
#> SRR1274567 3 0.6300 0.625 0.032 0.044 0.660 0.264
#> SRR1274568 3 0.5904 0.651 0.112 0.012 0.724 0.152
#> SRR1274569 1 0.2830 0.918 0.900 0.000 0.040 0.060
#> SRR1274570 1 0.1557 0.954 0.944 0.000 0.000 0.056
#> SRR1274571 1 0.1557 0.954 0.944 0.000 0.000 0.056
#> SRR1274572 1 0.0000 0.968 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0188 0.758 0.000 0.996 0.000 0.004
#> SRR1274574 2 0.3216 0.694 0.000 0.880 0.044 0.076
#> SRR1274575 4 0.6267 0.953 0.000 0.240 0.112 0.648
#> SRR1274576 2 0.5883 0.494 0.000 0.640 0.300 0.060
#> SRR1274577 3 0.2982 0.699 0.068 0.004 0.896 0.032
#> SRR1274578 1 0.0000 0.968 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.968 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.968 1.000 0.000 0.000 0.000
#> SRR1274581 3 0.7210 0.634 0.276 0.000 0.540 0.184
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.4509 0.62625 0.000 0.728 0.224 0.044 0.004
#> SRR1274529 4 0.2230 0.79283 0.000 0.116 0.000 0.884 0.000
#> SRR1274530 4 0.2230 0.79283 0.000 0.116 0.000 0.884 0.000
#> SRR1274531 5 0.7254 -0.02223 0.000 0.024 0.328 0.248 0.400
#> SRR1274532 2 0.0000 0.71273 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 3 0.6532 -0.04979 0.000 0.020 0.520 0.132 0.328
#> SRR1274534 3 0.6440 -0.05171 0.000 0.024 0.532 0.112 0.332
#> SRR1274535 3 0.6287 -0.05019 0.000 0.012 0.536 0.124 0.328
#> SRR1274536 4 0.5451 0.55471 0.000 0.024 0.104 0.700 0.172
#> SRR1274537 2 0.6692 0.46196 0.000 0.504 0.308 0.172 0.016
#> SRR1274538 2 0.5860 0.39740 0.000 0.556 0.068 0.360 0.016
#> SRR1274539 3 0.6440 -0.05171 0.000 0.024 0.532 0.112 0.332
#> SRR1274540 2 0.0162 0.71160 0.000 0.996 0.000 0.004 0.000
#> SRR1274541 2 0.0162 0.71160 0.000 0.996 0.000 0.004 0.000
#> SRR1274542 2 0.0000 0.71273 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.6112 0.46701 0.000 0.532 0.364 0.088 0.016
#> SRR1274544 2 0.0000 0.71273 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 4 0.3864 0.70617 0.000 0.188 0.020 0.784 0.008
#> SRR1274546 2 0.2278 0.69971 0.000 0.908 0.060 0.032 0.000
#> SRR1274547 2 0.5235 0.24719 0.000 0.524 0.024 0.440 0.012
#> SRR1274548 2 0.6544 0.48787 0.000 0.548 0.180 0.256 0.016
#> SRR1274549 3 0.6893 0.08910 0.028 0.000 0.468 0.152 0.352
#> SRR1274550 3 0.2481 0.07611 0.032 0.012 0.916 0.016 0.024
#> SRR1274551 2 0.0000 0.71273 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.2927 0.00456 0.000 0.000 0.868 0.092 0.040
#> SRR1274553 3 0.2026 0.06116 0.012 0.000 0.924 0.056 0.008
#> SRR1274554 3 0.6708 -0.07489 0.000 0.012 0.480 0.180 0.328
#> SRR1274555 3 0.5705 0.16750 0.036 0.000 0.488 0.024 0.452
#> SRR1274556 3 0.5977 0.15243 0.036 0.000 0.464 0.040 0.460
#> SRR1274557 3 0.6793 0.12585 0.032 0.012 0.524 0.096 0.336
#> SRR1274558 5 0.5716 0.10822 0.036 0.012 0.464 0.008 0.480
#> SRR1274559 5 0.6638 0.04019 0.256 0.000 0.252 0.004 0.488
#> SRR1274560 1 0.6771 -0.10844 0.368 0.000 0.272 0.000 0.360
#> SRR1274561 3 0.6666 0.11789 0.224 0.000 0.472 0.004 0.300
#> SRR1274562 3 0.6931 0.12126 0.236 0.000 0.448 0.012 0.304
#> SRR1274563 3 0.5690 0.16904 0.036 0.000 0.508 0.024 0.432
#> SRR1274564 4 0.4067 0.72364 0.000 0.024 0.060 0.816 0.100
#> SRR1274565 4 0.3587 0.74956 0.000 0.036 0.096 0.844 0.024
#> SRR1274566 4 0.3736 0.74260 0.000 0.024 0.064 0.840 0.072
#> SRR1274567 5 0.6904 -0.03736 0.000 0.012 0.204 0.380 0.404
#> SRR1274568 5 0.5716 0.10822 0.036 0.012 0.464 0.008 0.480
#> SRR1274569 1 0.1965 0.85086 0.904 0.000 0.000 0.000 0.096
#> SRR1274570 1 0.1792 0.85803 0.916 0.000 0.000 0.000 0.084
#> SRR1274571 1 0.1792 0.85803 0.916 0.000 0.000 0.000 0.084
#> SRR1274572 1 0.0324 0.86805 0.992 0.000 0.004 0.000 0.004
#> SRR1274573 2 0.0000 0.71273 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.4814 0.33506 0.000 0.568 0.016 0.412 0.004
#> SRR1274575 4 0.2230 0.79283 0.000 0.116 0.000 0.884 0.000
#> SRR1274576 2 0.7915 0.16513 0.000 0.364 0.360 0.104 0.172
#> SRR1274577 3 0.6175 -0.07857 0.008 0.004 0.568 0.116 0.304
#> SRR1274578 1 0.0290 0.86737 0.992 0.000 0.008 0.000 0.000
#> SRR1274579 1 0.0324 0.86805 0.992 0.000 0.004 0.000 0.004
#> SRR1274580 1 0.0324 0.86805 0.992 0.000 0.004 0.000 0.004
#> SRR1274581 3 0.7176 0.10746 0.252 0.000 0.412 0.020 0.316
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 6 0.5434 0.2319 0.000 0.456 0.000 0.060 0.024 0.460
#> SRR1274529 4 0.1387 0.6733 0.000 0.068 0.000 0.932 0.000 0.000
#> SRR1274530 4 0.1387 0.6733 0.000 0.068 0.000 0.932 0.000 0.000
#> SRR1274531 3 0.4520 0.6381 0.000 0.020 0.676 0.024 0.276 0.004
#> SRR1274532 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0692 0.9789 0.000 0.020 0.000 0.000 0.976 0.004
#> SRR1274534 5 0.0653 0.9746 0.000 0.004 0.004 0.012 0.980 0.000
#> SRR1274535 5 0.0458 0.9798 0.000 0.016 0.000 0.000 0.984 0.000
#> SRR1274536 3 0.5141 0.4788 0.000 0.004 0.608 0.304 0.076 0.008
#> SRR1274537 6 0.6106 0.3637 0.000 0.392 0.000 0.096 0.048 0.464
#> SRR1274538 6 0.6079 0.3443 0.000 0.392 0.000 0.112 0.036 0.460
#> SRR1274539 5 0.0653 0.9746 0.000 0.004 0.004 0.012 0.980 0.000
#> SRR1274540 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 6 0.6517 0.3752 0.000 0.360 0.004 0.092 0.080 0.464
#> SRR1274544 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 4 0.3640 0.4928 0.000 0.204 0.000 0.764 0.004 0.028
#> SRR1274546 2 0.4076 -0.1778 0.000 0.540 0.000 0.000 0.008 0.452
#> SRR1274547 2 0.6408 -0.0655 0.000 0.400 0.000 0.372 0.024 0.204
#> SRR1274548 6 0.6178 0.3685 0.000 0.392 0.000 0.088 0.060 0.460
#> SRR1274549 3 0.1340 0.7631 0.000 0.000 0.948 0.040 0.004 0.008
#> SRR1274550 3 0.2177 0.7591 0.000 0.008 0.908 0.000 0.052 0.032
#> SRR1274551 2 0.0000 0.7450 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 3 0.3602 0.7304 0.000 0.000 0.784 0.056 0.160 0.000
#> SRR1274553 3 0.2560 0.7623 0.000 0.000 0.872 0.036 0.092 0.000
#> SRR1274554 5 0.0692 0.9789 0.000 0.020 0.000 0.000 0.976 0.004
#> SRR1274555 3 0.1297 0.7586 0.000 0.000 0.948 0.012 0.000 0.040
#> SRR1274556 3 0.1708 0.7610 0.000 0.000 0.932 0.040 0.004 0.024
#> SRR1274557 3 0.1679 0.7588 0.000 0.012 0.936 0.000 0.016 0.036
#> SRR1274558 6 0.6011 -0.3629 0.000 0.000 0.336 0.000 0.248 0.416
#> SRR1274559 3 0.4951 0.5168 0.012 0.000 0.516 0.020 0.012 0.440
#> SRR1274560 3 0.5028 0.5174 0.016 0.000 0.516 0.020 0.012 0.436
#> SRR1274561 3 0.3717 0.7483 0.076 0.000 0.832 0.032 0.020 0.040
#> SRR1274562 3 0.3287 0.7500 0.076 0.000 0.852 0.016 0.012 0.044
#> SRR1274563 3 0.1297 0.7600 0.000 0.000 0.948 0.000 0.012 0.040
#> SRR1274564 4 0.5199 0.3811 0.000 0.004 0.320 0.592 0.076 0.008
#> SRR1274565 4 0.5310 0.5238 0.000 0.024 0.240 0.644 0.088 0.004
#> SRR1274566 4 0.5075 0.4600 0.000 0.004 0.288 0.624 0.076 0.008
#> SRR1274567 3 0.4873 0.5644 0.000 0.004 0.660 0.256 0.072 0.008
#> SRR1274568 6 0.6044 -0.3250 0.000 0.000 0.308 0.000 0.276 0.416
#> SRR1274569 1 0.2631 0.9003 0.840 0.000 0.008 0.000 0.000 0.152
#> SRR1274570 1 0.1957 0.9229 0.888 0.000 0.000 0.000 0.000 0.112
#> SRR1274571 1 0.1957 0.9229 0.888 0.000 0.000 0.000 0.000 0.112
#> SRR1274572 1 0.0000 0.9344 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0547 0.7280 0.000 0.980 0.000 0.020 0.000 0.000
#> SRR1274574 2 0.5927 -0.0220 0.000 0.448 0.000 0.324 0.000 0.228
#> SRR1274575 4 0.1387 0.6733 0.000 0.068 0.000 0.932 0.000 0.000
#> SRR1274576 6 0.6530 0.3586 0.000 0.336 0.000 0.072 0.124 0.468
#> SRR1274577 3 0.7278 0.2599 0.000 0.012 0.384 0.084 0.340 0.180
#> SRR1274578 1 0.1204 0.9289 0.944 0.000 0.000 0.000 0.000 0.056
#> SRR1274579 1 0.0000 0.9344 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.9344 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.4706 0.7121 0.080 0.000 0.744 0.064 0.000 0.112
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["SD", "NMF"]
# you can also extract it by
# res = res_list["SD:NMF"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'SD' method.
#> Subgroups are detected by 'NMF' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.999 0.962 0.984 0.4849 0.516 0.516
#> 3 3 0.893 0.901 0.960 0.2981 0.790 0.619
#> 4 4 0.737 0.682 0.816 0.1612 0.771 0.472
#> 5 5 0.707 0.701 0.811 0.0531 0.934 0.766
#> 6 6 0.835 0.776 0.892 0.0475 0.908 0.633
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.984 0.000 1.000
#> SRR1274529 2 0.0000 0.984 0.000 1.000
#> SRR1274530 2 0.0000 0.984 0.000 1.000
#> SRR1274531 2 0.1843 0.962 0.028 0.972
#> SRR1274532 2 0.0000 0.984 0.000 1.000
#> SRR1274533 2 0.0000 0.984 0.000 1.000
#> SRR1274534 2 0.0000 0.984 0.000 1.000
#> SRR1274535 2 0.0376 0.981 0.004 0.996
#> SRR1274536 2 0.0000 0.984 0.000 1.000
#> SRR1274537 2 0.0000 0.984 0.000 1.000
#> SRR1274538 2 0.0000 0.984 0.000 1.000
#> SRR1274539 2 0.0000 0.984 0.000 1.000
#> SRR1274540 2 0.0000 0.984 0.000 1.000
#> SRR1274541 2 0.0000 0.984 0.000 1.000
#> SRR1274542 2 0.0000 0.984 0.000 1.000
#> SRR1274543 2 0.0000 0.984 0.000 1.000
#> SRR1274544 2 0.0000 0.984 0.000 1.000
#> SRR1274545 2 0.0000 0.984 0.000 1.000
#> SRR1274546 2 0.0000 0.984 0.000 1.000
#> SRR1274547 2 0.0000 0.984 0.000 1.000
#> SRR1274548 2 0.0000 0.984 0.000 1.000
#> SRR1274549 1 0.0000 0.981 1.000 0.000
#> SRR1274550 1 0.0000 0.981 1.000 0.000
#> SRR1274551 2 0.0000 0.984 0.000 1.000
#> SRR1274552 1 0.0000 0.981 1.000 0.000
#> SRR1274553 1 0.0000 0.981 1.000 0.000
#> SRR1274554 2 0.1184 0.971 0.016 0.984
#> SRR1274555 1 0.0000 0.981 1.000 0.000
#> SRR1274556 1 0.0000 0.981 1.000 0.000
#> SRR1274557 1 0.0000 0.981 1.000 0.000
#> SRR1274558 2 0.4298 0.902 0.088 0.912
#> SRR1274559 1 0.0000 0.981 1.000 0.000
#> SRR1274560 1 0.0000 0.981 1.000 0.000
#> SRR1274561 1 0.0000 0.981 1.000 0.000
#> SRR1274562 1 0.0000 0.981 1.000 0.000
#> SRR1274563 1 0.0000 0.981 1.000 0.000
#> SRR1274564 2 0.0000 0.984 0.000 1.000
#> SRR1274565 2 0.0000 0.984 0.000 1.000
#> SRR1274566 2 0.0000 0.984 0.000 1.000
#> SRR1274567 1 0.9427 0.411 0.640 0.360
#> SRR1274568 2 0.8207 0.663 0.256 0.744
#> SRR1274569 1 0.0000 0.981 1.000 0.000
#> SRR1274570 1 0.0000 0.981 1.000 0.000
#> SRR1274571 1 0.0000 0.981 1.000 0.000
#> SRR1274572 1 0.0000 0.981 1.000 0.000
#> SRR1274573 2 0.0000 0.984 0.000 1.000
#> SRR1274574 2 0.0000 0.984 0.000 1.000
#> SRR1274575 2 0.0000 0.984 0.000 1.000
#> SRR1274576 2 0.0000 0.984 0.000 1.000
#> SRR1274577 2 0.5408 0.861 0.124 0.876
#> SRR1274578 1 0.0000 0.981 1.000 0.000
#> SRR1274579 1 0.0000 0.981 1.000 0.000
#> SRR1274580 1 0.0000 0.981 1.000 0.000
#> SRR1274581 1 0.0000 0.981 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274529 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274530 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274531 3 0.0000 0.936 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274533 3 0.6252 0.176 0.000 0.444 0.556
#> SRR1274534 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274535 3 0.0237 0.935 0.000 0.004 0.996
#> SRR1274536 3 0.0000 0.936 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274539 2 0.0892 0.936 0.000 0.980 0.020
#> SRR1274540 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274545 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274546 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274547 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274548 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.936 0.000 0.000 1.000
#> SRR1274550 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274551 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274552 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274553 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274554 3 0.4062 0.760 0.000 0.164 0.836
#> SRR1274555 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274556 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274557 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274558 2 0.6180 0.333 0.416 0.584 0.000
#> SRR1274559 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274560 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274561 1 0.3879 0.819 0.848 0.000 0.152
#> SRR1274562 1 0.0424 0.977 0.992 0.000 0.008
#> SRR1274563 3 0.0237 0.937 0.004 0.000 0.996
#> SRR1274564 2 0.5098 0.660 0.000 0.752 0.248
#> SRR1274565 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274566 2 0.4002 0.789 0.000 0.840 0.160
#> SRR1274567 3 0.0000 0.936 0.000 0.000 1.000
#> SRR1274568 2 0.6154 0.353 0.408 0.592 0.000
#> SRR1274569 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274575 2 0.0237 0.950 0.000 0.996 0.004
#> SRR1274576 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274577 2 0.0000 0.951 0.000 1.000 0.000
#> SRR1274578 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.984 1.000 0.000 0.000
#> SRR1274581 3 0.2066 0.891 0.060 0.000 0.940
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.4761 0.7575 0.000 0.628 0.000 0.372
#> SRR1274529 4 0.0000 0.6953 0.000 0.000 0.000 1.000
#> SRR1274530 4 0.0000 0.6953 0.000 0.000 0.000 1.000
#> SRR1274531 3 0.1356 0.8521 0.000 0.032 0.960 0.008
#> SRR1274532 2 0.4804 0.7544 0.000 0.616 0.000 0.384
#> SRR1274533 2 0.3161 0.4810 0.000 0.864 0.124 0.012
#> SRR1274534 2 0.1302 0.5984 0.000 0.956 0.000 0.044
#> SRR1274535 2 0.3751 0.3495 0.000 0.800 0.196 0.004
#> SRR1274536 4 0.4790 0.3938 0.000 0.000 0.380 0.620
#> SRR1274537 2 0.4713 0.7568 0.000 0.640 0.000 0.360
#> SRR1274538 2 0.4898 0.7277 0.000 0.584 0.000 0.416
#> SRR1274539 2 0.1452 0.5931 0.000 0.956 0.008 0.036
#> SRR1274540 2 0.4916 0.7181 0.000 0.576 0.000 0.424
#> SRR1274541 2 0.4804 0.7544 0.000 0.616 0.000 0.384
#> SRR1274542 2 0.4804 0.7544 0.000 0.616 0.000 0.384
#> SRR1274543 2 0.4761 0.7575 0.000 0.628 0.000 0.372
#> SRR1274544 2 0.4830 0.7492 0.000 0.608 0.000 0.392
#> SRR1274545 4 0.0188 0.6919 0.000 0.004 0.000 0.996
#> SRR1274546 2 0.4776 0.7567 0.000 0.624 0.000 0.376
#> SRR1274547 4 0.1474 0.6302 0.000 0.052 0.000 0.948
#> SRR1274548 2 0.4713 0.7569 0.000 0.640 0.000 0.360
#> SRR1274549 3 0.0188 0.8548 0.000 0.000 0.996 0.004
#> SRR1274550 3 0.4406 0.7306 0.000 0.300 0.700 0.000
#> SRR1274551 2 0.4933 0.7076 0.000 0.568 0.000 0.432
#> SRR1274552 3 0.4624 0.7039 0.000 0.340 0.660 0.000
#> SRR1274553 3 0.4585 0.7107 0.000 0.332 0.668 0.000
#> SRR1274554 2 0.4283 0.2383 0.000 0.740 0.256 0.004
#> SRR1274555 3 0.0188 0.8591 0.000 0.004 0.996 0.000
#> SRR1274556 3 0.0000 0.8572 0.000 0.000 1.000 0.000
#> SRR1274557 3 0.0188 0.8591 0.000 0.004 0.996 0.000
#> SRR1274558 1 0.6155 0.0163 0.536 0.412 0.000 0.052
#> SRR1274559 1 0.0000 0.9046 1.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.9046 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.3649 0.7049 0.796 0.000 0.204 0.000
#> SRR1274562 1 0.2216 0.8431 0.908 0.000 0.092 0.000
#> SRR1274563 3 0.0188 0.8591 0.000 0.004 0.996 0.000
#> SRR1274564 4 0.4477 0.5071 0.000 0.000 0.312 0.688
#> SRR1274565 4 0.0376 0.6945 0.000 0.004 0.004 0.992
#> SRR1274566 4 0.4331 0.5369 0.000 0.000 0.288 0.712
#> SRR1274567 4 0.4790 0.3938 0.000 0.000 0.380 0.620
#> SRR1274568 2 0.5666 0.3938 0.348 0.616 0.000 0.036
#> SRR1274569 1 0.0000 0.9046 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.9046 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.9046 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0707 0.9027 0.980 0.020 0.000 0.000
#> SRR1274573 2 0.4877 0.7359 0.000 0.592 0.000 0.408
#> SRR1274574 4 0.4977 -0.5597 0.000 0.460 0.000 0.540
#> SRR1274575 4 0.0000 0.6953 0.000 0.000 0.000 1.000
#> SRR1274576 2 0.4431 0.7360 0.000 0.696 0.000 0.304
#> SRR1274577 2 0.0336 0.5767 0.000 0.992 0.000 0.008
#> SRR1274578 1 0.1302 0.8976 0.956 0.044 0.000 0.000
#> SRR1274579 1 0.1302 0.8976 0.956 0.044 0.000 0.000
#> SRR1274580 1 0.1302 0.8976 0.956 0.044 0.000 0.000
#> SRR1274581 3 0.3110 0.8049 0.056 0.048 0.892 0.004
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0290 0.8417 0.000 0.992 0.000 0.000 0.008
#> SRR1274529 4 0.2605 0.8383 0.000 0.148 0.000 0.852 0.000
#> SRR1274530 4 0.2605 0.8383 0.000 0.148 0.000 0.852 0.000
#> SRR1274531 3 0.1341 0.8151 0.000 0.000 0.944 0.000 0.056
#> SRR1274532 2 0.0609 0.8472 0.000 0.980 0.000 0.020 0.000
#> SRR1274533 5 0.5459 0.5226 0.000 0.316 0.084 0.000 0.600
#> SRR1274534 2 0.4291 0.0312 0.000 0.536 0.000 0.000 0.464
#> SRR1274535 5 0.5791 0.6028 0.000 0.260 0.140 0.000 0.600
#> SRR1274536 4 0.2773 0.7432 0.000 0.000 0.164 0.836 0.000
#> SRR1274537 2 0.0703 0.8330 0.000 0.976 0.000 0.000 0.024
#> SRR1274538 2 0.0794 0.8451 0.000 0.972 0.000 0.028 0.000
#> SRR1274539 2 0.4268 0.1026 0.000 0.556 0.000 0.000 0.444
#> SRR1274540 2 0.0794 0.8451 0.000 0.972 0.000 0.028 0.000
#> SRR1274541 2 0.0609 0.8472 0.000 0.980 0.000 0.020 0.000
#> SRR1274542 2 0.0609 0.8472 0.000 0.980 0.000 0.020 0.000
#> SRR1274543 2 0.0162 0.8455 0.000 0.996 0.000 0.004 0.000
#> SRR1274544 2 0.0609 0.8472 0.000 0.980 0.000 0.020 0.000
#> SRR1274545 4 0.2929 0.8224 0.000 0.180 0.000 0.820 0.000
#> SRR1274546 2 0.0000 0.8445 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 4 0.4030 0.6217 0.000 0.352 0.000 0.648 0.000
#> SRR1274548 2 0.0404 0.8400 0.000 0.988 0.000 0.000 0.012
#> SRR1274549 3 0.1907 0.8100 0.000 0.000 0.928 0.044 0.028
#> SRR1274550 5 0.4249 0.5707 0.000 0.000 0.432 0.000 0.568
#> SRR1274551 2 0.0794 0.8451 0.000 0.972 0.000 0.028 0.000
#> SRR1274552 5 0.4436 0.6235 0.000 0.008 0.396 0.000 0.596
#> SRR1274553 5 0.4192 0.6172 0.000 0.000 0.404 0.000 0.596
#> SRR1274554 2 0.6163 0.1446 0.000 0.536 0.300 0.000 0.164
#> SRR1274555 3 0.0000 0.8616 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0290 0.8578 0.000 0.000 0.992 0.000 0.008
#> SRR1274557 3 0.0000 0.8616 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 1 0.4225 0.2970 0.632 0.364 0.000 0.004 0.000
#> SRR1274559 1 0.0162 0.7851 0.996 0.000 0.000 0.004 0.000
#> SRR1274560 1 0.0162 0.7851 0.996 0.000 0.000 0.004 0.000
#> SRR1274561 1 0.2286 0.7213 0.888 0.000 0.108 0.004 0.000
#> SRR1274562 1 0.4828 0.4754 0.648 0.000 0.320 0.012 0.020
#> SRR1274563 3 0.0000 0.8616 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.3115 0.7966 0.000 0.036 0.112 0.852 0.000
#> SRR1274565 4 0.6062 0.6014 0.000 0.308 0.108 0.572 0.012
#> SRR1274566 4 0.3371 0.7995 0.000 0.040 0.104 0.848 0.008
#> SRR1274567 4 0.3093 0.7384 0.000 0.000 0.168 0.824 0.008
#> SRR1274568 2 0.4479 0.5533 0.264 0.700 0.000 0.000 0.036
#> SRR1274569 1 0.0162 0.7851 0.996 0.000 0.000 0.004 0.000
#> SRR1274570 1 0.0000 0.7853 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.7853 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.4049 0.7400 0.792 0.000 0.000 0.084 0.124
#> SRR1274573 2 0.0794 0.8451 0.000 0.972 0.000 0.028 0.000
#> SRR1274574 2 0.2127 0.7654 0.000 0.892 0.000 0.108 0.000
#> SRR1274575 4 0.2605 0.8383 0.000 0.148 0.000 0.852 0.000
#> SRR1274576 2 0.1043 0.8215 0.000 0.960 0.000 0.000 0.040
#> SRR1274577 2 0.5739 0.3602 0.000 0.556 0.000 0.100 0.344
#> SRR1274578 1 0.5820 0.6377 0.572 0.000 0.000 0.120 0.308
#> SRR1274579 1 0.5788 0.6434 0.580 0.000 0.000 0.120 0.300
#> SRR1274580 1 0.5772 0.6459 0.584 0.000 0.000 0.120 0.296
#> SRR1274581 3 0.6376 0.3458 0.040 0.000 0.496 0.068 0.396
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.0632 0.8582 0.000 0.024 0.000 0.976 0.000 0.000
#> SRR1274530 4 0.0632 0.8582 0.000 0.024 0.000 0.976 0.000 0.000
#> SRR1274531 3 0.0935 0.7527 0.000 0.000 0.964 0.000 0.032 0.004
#> SRR1274532 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.2250 0.8629 0.000 0.040 0.064 0.000 0.896 0.000
#> SRR1274534 5 0.2912 0.7199 0.000 0.216 0.000 0.000 0.784 0.000
#> SRR1274535 5 0.2147 0.8686 0.000 0.020 0.084 0.000 0.896 0.000
#> SRR1274536 4 0.0508 0.8500 0.000 0.000 0.012 0.984 0.000 0.004
#> SRR1274537 2 0.0146 0.9777 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274538 2 0.0146 0.9768 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274539 5 0.3189 0.6951 0.000 0.236 0.004 0.000 0.760 0.000
#> SRR1274540 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0146 0.9777 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274544 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 4 0.1327 0.8347 0.000 0.064 0.000 0.936 0.000 0.000
#> SRR1274546 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 4 0.3409 0.5676 0.000 0.300 0.000 0.700 0.000 0.000
#> SRR1274548 2 0.0146 0.9777 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274549 3 0.1888 0.7198 0.000 0.000 0.916 0.068 0.012 0.004
#> SRR1274550 5 0.2100 0.8574 0.000 0.000 0.112 0.004 0.884 0.000
#> SRR1274551 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.2006 0.8629 0.000 0.000 0.104 0.004 0.892 0.000
#> SRR1274553 5 0.2006 0.8629 0.000 0.000 0.104 0.004 0.892 0.000
#> SRR1274554 3 0.5543 0.0701 0.000 0.452 0.464 0.004 0.048 0.032
#> SRR1274555 3 0.0363 0.7599 0.000 0.000 0.988 0.000 0.012 0.000
#> SRR1274556 3 0.0291 0.7587 0.000 0.000 0.992 0.000 0.004 0.004
#> SRR1274557 3 0.0260 0.7598 0.000 0.000 0.992 0.000 0.008 0.000
#> SRR1274558 1 0.2996 0.5367 0.772 0.228 0.000 0.000 0.000 0.000
#> SRR1274559 1 0.0000 0.8346 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.8346 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.1556 0.7696 0.920 0.000 0.080 0.000 0.000 0.000
#> SRR1274562 3 0.4591 0.0755 0.452 0.000 0.516 0.000 0.004 0.028
#> SRR1274563 3 0.0363 0.7596 0.000 0.000 0.988 0.000 0.012 0.000
#> SRR1274564 4 0.0551 0.8516 0.000 0.004 0.008 0.984 0.000 0.004
#> SRR1274565 4 0.7122 0.1300 0.000 0.284 0.264 0.396 0.028 0.028
#> SRR1274566 4 0.1508 0.8453 0.000 0.004 0.016 0.948 0.020 0.012
#> SRR1274567 4 0.2063 0.8209 0.000 0.000 0.060 0.912 0.020 0.008
#> SRR1274568 2 0.3865 0.6728 0.208 0.756 0.004 0.000 0.016 0.016
#> SRR1274569 1 0.0363 0.8355 0.988 0.000 0.000 0.000 0.000 0.012
#> SRR1274570 1 0.0363 0.8355 0.988 0.000 0.000 0.000 0.000 0.012
#> SRR1274571 1 0.0363 0.8355 0.988 0.000 0.000 0.000 0.000 0.012
#> SRR1274572 1 0.3860 -0.1716 0.528 0.000 0.000 0.000 0.000 0.472
#> SRR1274573 2 0.0000 0.9796 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0260 0.9730 0.000 0.992 0.000 0.008 0.000 0.000
#> SRR1274575 4 0.0632 0.8582 0.000 0.024 0.000 0.976 0.000 0.000
#> SRR1274576 2 0.0260 0.9746 0.000 0.992 0.000 0.000 0.008 0.000
#> SRR1274577 6 0.4113 0.6118 0.000 0.188 0.004 0.004 0.056 0.748
#> SRR1274578 6 0.2527 0.8430 0.168 0.000 0.000 0.000 0.000 0.832
#> SRR1274579 6 0.2631 0.8436 0.180 0.000 0.000 0.000 0.000 0.820
#> SRR1274580 6 0.2730 0.8342 0.192 0.000 0.000 0.000 0.000 0.808
#> SRR1274581 3 0.7174 0.2307 0.068 0.000 0.428 0.020 0.164 0.320
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "hclust"]
# you can also extract it by
# res = res_list["CV:hclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'hclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 3.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.398 0.743 0.859 0.2917 0.669 0.669
#> 3 3 0.656 0.773 0.910 0.5074 0.915 0.874
#> 4 4 0.595 0.654 0.851 0.1852 0.876 0.796
#> 5 5 0.604 0.667 0.874 0.0894 0.904 0.811
#> 6 6 0.685 0.668 0.845 0.0525 0.933 0.860
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 3
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.8854 0.000 1.000
#> SRR1274529 2 0.0000 0.8854 0.000 1.000
#> SRR1274530 2 0.0000 0.8854 0.000 1.000
#> SRR1274531 2 0.2603 0.8556 0.044 0.956
#> SRR1274532 2 0.0000 0.8854 0.000 1.000
#> SRR1274533 2 0.6048 0.7243 0.148 0.852
#> SRR1274534 2 0.1414 0.8758 0.020 0.980
#> SRR1274535 2 0.1414 0.8758 0.020 0.980
#> SRR1274536 2 0.3114 0.8445 0.056 0.944
#> SRR1274537 2 0.0000 0.8854 0.000 1.000
#> SRR1274538 2 0.0000 0.8854 0.000 1.000
#> SRR1274539 2 0.1414 0.8758 0.020 0.980
#> SRR1274540 2 0.0000 0.8854 0.000 1.000
#> SRR1274541 2 0.0000 0.8854 0.000 1.000
#> SRR1274542 2 0.0000 0.8854 0.000 1.000
#> SRR1274543 2 0.0000 0.8854 0.000 1.000
#> SRR1274544 2 0.0000 0.8854 0.000 1.000
#> SRR1274545 2 0.0000 0.8854 0.000 1.000
#> SRR1274546 2 0.0000 0.8854 0.000 1.000
#> SRR1274547 2 0.0000 0.8854 0.000 1.000
#> SRR1274548 2 0.0000 0.8854 0.000 1.000
#> SRR1274549 2 0.3114 0.8445 0.056 0.944
#> SRR1274550 2 0.7674 0.5905 0.224 0.776
#> SRR1274551 2 0.0000 0.8854 0.000 1.000
#> SRR1274552 2 0.7674 0.5905 0.224 0.776
#> SRR1274553 2 0.7674 0.5905 0.224 0.776
#> SRR1274554 2 0.1414 0.8758 0.020 0.980
#> SRR1274555 1 0.9358 0.5707 0.648 0.352
#> SRR1274556 2 0.9732 -0.0533 0.404 0.596
#> SRR1274557 1 0.9358 0.5707 0.648 0.352
#> SRR1274558 2 0.6623 0.6335 0.172 0.828
#> SRR1274559 2 0.7883 0.4752 0.236 0.764
#> SRR1274560 2 0.7883 0.4752 0.236 0.764
#> SRR1274561 1 0.8909 0.6528 0.692 0.308
#> SRR1274562 1 0.9170 0.6314 0.668 0.332
#> SRR1274563 1 0.9358 0.5707 0.648 0.352
#> SRR1274564 2 0.3114 0.8445 0.056 0.944
#> SRR1274565 2 0.0000 0.8854 0.000 1.000
#> SRR1274566 2 0.0376 0.8837 0.004 0.996
#> SRR1274567 2 0.0376 0.8837 0.004 0.996
#> SRR1274568 2 0.5059 0.7533 0.112 0.888
#> SRR1274569 1 0.9754 0.6734 0.592 0.408
#> SRR1274570 1 0.9661 0.6936 0.608 0.392
#> SRR1274571 1 0.9661 0.6936 0.608 0.392
#> SRR1274572 1 0.9661 0.6936 0.608 0.392
#> SRR1274573 2 0.0000 0.8854 0.000 1.000
#> SRR1274574 2 0.0000 0.8854 0.000 1.000
#> SRR1274575 2 0.0000 0.8854 0.000 1.000
#> SRR1274576 2 0.0000 0.8854 0.000 1.000
#> SRR1274577 2 0.2043 0.8704 0.032 0.968
#> SRR1274578 2 0.9963 -0.4455 0.464 0.536
#> SRR1274579 1 0.9661 0.6936 0.608 0.392
#> SRR1274580 1 0.9661 0.6936 0.608 0.392
#> SRR1274581 2 0.8016 0.5094 0.244 0.756
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274531 2 0.1860 0.8921 0.000 0.948 0.052
#> SRR1274532 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274533 2 0.4326 0.7912 0.012 0.844 0.144
#> SRR1274534 2 0.0892 0.9105 0.000 0.980 0.020
#> SRR1274535 2 0.0892 0.9105 0.000 0.980 0.020
#> SRR1274536 2 0.1964 0.8880 0.000 0.944 0.056
#> SRR1274537 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274539 2 0.0892 0.9105 0.000 0.980 0.020
#> SRR1274540 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274549 2 0.1964 0.8880 0.000 0.944 0.056
#> SRR1274550 2 0.6852 0.5173 0.036 0.664 0.300
#> SRR1274551 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274552 2 0.6852 0.5173 0.036 0.664 0.300
#> SRR1274553 2 0.6852 0.5173 0.036 0.664 0.300
#> SRR1274554 2 0.0892 0.9105 0.000 0.980 0.020
#> SRR1274555 3 0.1529 0.6139 0.040 0.000 0.960
#> SRR1274556 3 0.7174 -0.0731 0.024 0.460 0.516
#> SRR1274557 3 0.1529 0.6139 0.040 0.000 0.960
#> SRR1274558 2 0.4796 0.6925 0.220 0.780 0.000
#> SRR1274559 2 0.5678 0.5502 0.316 0.684 0.000
#> SRR1274560 2 0.5678 0.5502 0.316 0.684 0.000
#> SRR1274561 1 0.7671 0.0297 0.492 0.044 0.464
#> SRR1274562 3 0.6204 -0.0153 0.424 0.000 0.576
#> SRR1274563 3 0.1860 0.6090 0.052 0.000 0.948
#> SRR1274564 2 0.1964 0.8880 0.000 0.944 0.056
#> SRR1274565 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274566 2 0.0237 0.9166 0.000 0.996 0.004
#> SRR1274567 2 0.0237 0.9166 0.000 0.996 0.004
#> SRR1274568 2 0.4110 0.7807 0.152 0.844 0.004
#> SRR1274569 1 0.3267 0.7160 0.884 0.116 0.000
#> SRR1274570 1 0.0592 0.8332 0.988 0.012 0.000
#> SRR1274571 1 0.0592 0.8332 0.988 0.012 0.000
#> SRR1274572 1 0.0592 0.8332 0.988 0.012 0.000
#> SRR1274573 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.9178 0.000 1.000 0.000
#> SRR1274577 2 0.1482 0.9069 0.012 0.968 0.020
#> SRR1274578 1 0.4399 0.5708 0.812 0.188 0.000
#> SRR1274579 1 0.0592 0.8332 0.988 0.012 0.000
#> SRR1274580 1 0.0592 0.8332 0.988 0.012 0.000
#> SRR1274581 2 0.8380 0.3737 0.276 0.600 0.124
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274531 2 0.4761 0.46281 0.000 0.664 0.332 0.004
#> SRR1274532 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274533 2 0.4972 -0.01659 0.000 0.544 0.456 0.000
#> SRR1274534 2 0.0817 0.87398 0.000 0.976 0.024 0.000
#> SRR1274535 2 0.0921 0.87228 0.000 0.972 0.028 0.000
#> SRR1274536 2 0.4730 0.40237 0.000 0.636 0.364 0.000
#> SRR1274537 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274539 2 0.0817 0.87398 0.000 0.976 0.024 0.000
#> SRR1274540 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274549 2 0.4730 0.40237 0.000 0.636 0.364 0.000
#> SRR1274550 3 0.4730 0.46209 0.000 0.364 0.636 0.000
#> SRR1274551 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.4730 0.46209 0.000 0.364 0.636 0.000
#> SRR1274553 3 0.4730 0.46209 0.000 0.364 0.636 0.000
#> SRR1274554 2 0.1211 0.86566 0.000 0.960 0.040 0.000
#> SRR1274555 3 0.4955 -0.18963 0.000 0.000 0.556 0.444
#> SRR1274556 3 0.3471 0.21313 0.000 0.072 0.868 0.060
#> SRR1274557 3 0.4955 -0.18963 0.000 0.000 0.556 0.444
#> SRR1274558 2 0.3873 0.63200 0.228 0.772 0.000 0.000
#> SRR1274559 2 0.4543 0.46842 0.324 0.676 0.000 0.000
#> SRR1274560 2 0.4543 0.46842 0.324 0.676 0.000 0.000
#> SRR1274561 4 0.6854 -0.00644 0.460 0.044 0.028 0.468
#> SRR1274562 4 0.7207 0.18921 0.376 0.000 0.144 0.480
#> SRR1274563 4 0.5168 -0.14940 0.004 0.000 0.492 0.504
#> SRR1274564 2 0.4730 0.40237 0.000 0.636 0.364 0.000
#> SRR1274565 2 0.0817 0.87351 0.000 0.976 0.024 0.000
#> SRR1274566 2 0.1867 0.83409 0.000 0.928 0.072 0.000
#> SRR1274567 2 0.1867 0.83409 0.000 0.928 0.072 0.000
#> SRR1274568 2 0.3355 0.72443 0.160 0.836 0.004 0.000
#> SRR1274569 1 0.2408 0.72580 0.896 0.104 0.000 0.000
#> SRR1274570 1 0.0000 0.88288 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.88288 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.88288 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274576 2 0.0000 0.88476 0.000 1.000 0.000 0.000
#> SRR1274577 2 0.1284 0.87009 0.012 0.964 0.024 0.000
#> SRR1274578 1 0.3486 0.51832 0.812 0.188 0.000 0.000
#> SRR1274579 1 0.0000 0.88288 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.88288 1.000 0.000 0.000 0.000
#> SRR1274581 4 0.7953 -0.15882 0.148 0.412 0.024 0.416
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274530 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274531 2 0.497 -0.184 0.000 0.540 0.008 0.436 0.016
#> SRR1274532 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 4 0.389 0.389 0.000 0.320 0.000 0.680 0.000
#> SRR1274534 2 0.120 0.866 0.000 0.952 0.000 0.048 0.000
#> SRR1274535 2 0.141 0.858 0.000 0.940 0.000 0.060 0.000
#> SRR1274536 4 0.475 0.232 0.000 0.492 0.000 0.492 0.016
#> SRR1274537 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.120 0.866 0.000 0.952 0.000 0.048 0.000
#> SRR1274540 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 2 0.475 -0.361 0.000 0.492 0.000 0.492 0.016
#> SRR1274550 4 0.444 0.466 0.000 0.136 0.104 0.760 0.000
#> SRR1274551 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 4 0.444 0.466 0.000 0.136 0.104 0.760 0.000
#> SRR1274553 4 0.444 0.466 0.000 0.136 0.104 0.760 0.000
#> SRR1274554 2 0.161 0.850 0.000 0.928 0.000 0.072 0.000
#> SRR1274555 3 0.000 0.785 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 4 0.488 -0.197 0.000 0.008 0.392 0.584 0.016
#> SRR1274557 3 0.000 0.785 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 2 0.334 0.608 0.228 0.772 0.000 0.000 0.000
#> SRR1274559 2 0.391 0.398 0.324 0.676 0.000 0.000 0.000
#> SRR1274560 2 0.391 0.398 0.324 0.676 0.000 0.000 0.000
#> SRR1274561 1 0.828 -0.239 0.368 0.044 0.308 0.036 0.244
#> SRR1274562 3 0.740 0.287 0.280 0.000 0.436 0.040 0.244
#> SRR1274563 3 0.172 0.775 0.004 0.000 0.936 0.008 0.052
#> SRR1274564 4 0.475 0.232 0.000 0.492 0.000 0.492 0.016
#> SRR1274565 2 0.112 0.864 0.000 0.956 0.000 0.044 0.000
#> SRR1274566 2 0.191 0.810 0.000 0.908 0.000 0.092 0.000
#> SRR1274567 2 0.191 0.810 0.000 0.908 0.000 0.092 0.000
#> SRR1274568 2 0.301 0.714 0.160 0.832 0.000 0.008 0.000
#> SRR1274569 1 0.207 0.699 0.896 0.104 0.000 0.000 0.000
#> SRR1274570 1 0.000 0.821 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.000 0.821 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.000 0.821 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274576 2 0.000 0.894 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 2 0.181 0.858 0.012 0.936 0.000 0.044 0.008
#> SRR1274578 1 0.321 0.518 0.812 0.180 0.000 0.000 0.008
#> SRR1274579 1 0.000 0.821 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.000 0.821 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 5 0.356 0.000 0.000 0.260 0.000 0.000 0.740
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274530 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274531 2 0.7117 -0.0388 0.000 0.456 0.100 0.112 0.312 0.020
#> SRR1274532 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.3221 0.4972 0.000 0.264 0.000 0.000 0.736 0.000
#> SRR1274534 2 0.1075 0.8168 0.000 0.952 0.000 0.000 0.048 0.000
#> SRR1274535 2 0.1327 0.8086 0.000 0.936 0.000 0.000 0.064 0.000
#> SRR1274536 2 0.7568 -0.2266 0.000 0.380 0.108 0.176 0.316 0.020
#> SRR1274537 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 2 0.1075 0.8168 0.000 0.952 0.000 0.000 0.048 0.000
#> SRR1274540 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 2 0.7568 -0.2266 0.000 0.380 0.108 0.176 0.316 0.020
#> SRR1274550 5 0.1556 0.8380 0.000 0.080 0.000 0.000 0.920 0.000
#> SRR1274551 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.1556 0.8380 0.000 0.080 0.000 0.000 0.920 0.000
#> SRR1274553 5 0.1556 0.8380 0.000 0.080 0.000 0.000 0.920 0.000
#> SRR1274554 2 0.1531 0.8061 0.000 0.928 0.004 0.000 0.068 0.000
#> SRR1274555 3 0.3620 0.6654 0.000 0.000 0.648 0.352 0.000 0.000
#> SRR1274556 3 0.3761 0.2128 0.000 0.008 0.744 0.000 0.228 0.020
#> SRR1274557 3 0.3620 0.6654 0.000 0.000 0.648 0.352 0.000 0.000
#> SRR1274558 2 0.4586 0.5262 0.104 0.712 0.000 0.000 0.008 0.176
#> SRR1274559 2 0.5384 0.3263 0.200 0.616 0.000 0.000 0.008 0.176
#> SRR1274560 2 0.5384 0.3263 0.200 0.616 0.000 0.000 0.008 0.176
#> SRR1274561 4 0.4173 0.7218 0.268 0.044 0.000 0.688 0.000 0.000
#> SRR1274562 4 0.2597 0.6809 0.176 0.000 0.000 0.824 0.000 0.000
#> SRR1274563 3 0.3789 0.6051 0.000 0.000 0.584 0.416 0.000 0.000
#> SRR1274564 2 0.7568 -0.2266 0.000 0.380 0.108 0.176 0.316 0.020
#> SRR1274565 2 0.1511 0.8103 0.000 0.944 0.012 0.012 0.032 0.000
#> SRR1274566 2 0.3159 0.7237 0.000 0.856 0.056 0.056 0.032 0.000
#> SRR1274567 2 0.3159 0.7237 0.000 0.856 0.056 0.056 0.032 0.000
#> SRR1274568 2 0.4044 0.6233 0.064 0.772 0.000 0.000 0.016 0.148
#> SRR1274569 1 0.3566 0.6435 0.800 0.104 0.000 0.000 0.000 0.096
#> SRR1274570 1 0.1714 0.7819 0.908 0.000 0.000 0.000 0.000 0.092
#> SRR1274571 1 0.0000 0.8339 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.8339 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274576 2 0.0000 0.8424 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 2 0.1549 0.8090 0.000 0.936 0.000 0.000 0.044 0.020
#> SRR1274578 1 0.3156 0.5316 0.800 0.180 0.000 0.000 0.000 0.020
#> SRR1274579 1 0.0363 0.8310 0.988 0.000 0.000 0.000 0.000 0.012
#> SRR1274580 1 0.0363 0.8310 0.988 0.000 0.000 0.000 0.000 0.012
#> SRR1274581 6 0.2883 0.0000 0.000 0.212 0.000 0.000 0.000 0.788
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "kmeans"]
# you can also extract it by
# res = res_list["CV:kmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'kmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.549 0.882 0.927 0.4335 0.535 0.535
#> 3 3 0.658 0.828 0.905 0.3443 0.855 0.737
#> 4 4 0.619 0.640 0.821 0.1648 0.886 0.743
#> 5 5 0.663 0.572 0.775 0.0763 0.893 0.707
#> 6 6 0.707 0.508 0.767 0.0567 0.920 0.732
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.961 0.000 1.000
#> SRR1274529 2 0.000 0.961 0.000 1.000
#> SRR1274530 2 0.000 0.961 0.000 1.000
#> SRR1274531 2 0.680 0.766 0.180 0.820
#> SRR1274532 2 0.000 0.961 0.000 1.000
#> SRR1274533 2 0.295 0.914 0.052 0.948
#> SRR1274534 2 0.000 0.961 0.000 1.000
#> SRR1274535 2 0.000 0.961 0.000 1.000
#> SRR1274536 2 0.552 0.826 0.128 0.872
#> SRR1274537 2 0.000 0.961 0.000 1.000
#> SRR1274538 2 0.000 0.961 0.000 1.000
#> SRR1274539 2 0.000 0.961 0.000 1.000
#> SRR1274540 2 0.000 0.961 0.000 1.000
#> SRR1274541 2 0.000 0.961 0.000 1.000
#> SRR1274542 2 0.000 0.961 0.000 1.000
#> SRR1274543 2 0.000 0.961 0.000 1.000
#> SRR1274544 2 0.000 0.961 0.000 1.000
#> SRR1274545 2 0.000 0.961 0.000 1.000
#> SRR1274546 2 0.000 0.961 0.000 1.000
#> SRR1274547 2 0.000 0.961 0.000 1.000
#> SRR1274548 2 0.000 0.961 0.000 1.000
#> SRR1274549 1 0.706 0.776 0.808 0.192
#> SRR1274550 1 0.795 0.761 0.760 0.240
#> SRR1274551 2 0.000 0.961 0.000 1.000
#> SRR1274552 1 0.802 0.759 0.756 0.244
#> SRR1274553 1 0.802 0.759 0.756 0.244
#> SRR1274554 2 0.000 0.961 0.000 1.000
#> SRR1274555 1 0.000 0.826 1.000 0.000
#> SRR1274556 1 0.541 0.819 0.876 0.124
#> SRR1274557 1 0.430 0.829 0.912 0.088
#> SRR1274558 2 0.260 0.915 0.044 0.956
#> SRR1274559 1 0.706 0.864 0.808 0.192
#> SRR1274560 1 0.706 0.864 0.808 0.192
#> SRR1274561 1 0.000 0.826 1.000 0.000
#> SRR1274562 1 0.000 0.826 1.000 0.000
#> SRR1274563 1 0.000 0.826 1.000 0.000
#> SRR1274564 2 0.311 0.910 0.056 0.944
#> SRR1274565 2 0.000 0.961 0.000 1.000
#> SRR1274566 2 0.000 0.961 0.000 1.000
#> SRR1274567 2 0.584 0.815 0.140 0.860
#> SRR1274568 2 0.204 0.931 0.032 0.968
#> SRR1274569 1 0.706 0.864 0.808 0.192
#> SRR1274570 1 0.706 0.864 0.808 0.192
#> SRR1274571 1 0.706 0.864 0.808 0.192
#> SRR1274572 1 0.706 0.864 0.808 0.192
#> SRR1274573 2 0.000 0.961 0.000 1.000
#> SRR1274574 2 0.000 0.961 0.000 1.000
#> SRR1274575 2 0.000 0.961 0.000 1.000
#> SRR1274576 2 0.000 0.961 0.000 1.000
#> SRR1274577 2 0.000 0.961 0.000 1.000
#> SRR1274578 1 0.706 0.864 0.808 0.192
#> SRR1274579 1 0.706 0.864 0.808 0.192
#> SRR1274580 1 0.706 0.864 0.808 0.192
#> SRR1274581 2 0.990 -0.163 0.440 0.560
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274529 2 0.3528 0.864 0.016 0.892 0.092
#> SRR1274530 2 0.3528 0.864 0.016 0.892 0.092
#> SRR1274531 3 0.2165 0.796 0.000 0.064 0.936
#> SRR1274532 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274533 2 0.6062 0.364 0.000 0.616 0.384
#> SRR1274534 2 0.3038 0.859 0.000 0.896 0.104
#> SRR1274535 2 0.3752 0.820 0.000 0.856 0.144
#> SRR1274536 3 0.6096 0.570 0.016 0.280 0.704
#> SRR1274537 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274539 2 0.3038 0.859 0.000 0.896 0.104
#> SRR1274540 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274545 2 0.0747 0.912 0.016 0.984 0.000
#> SRR1274546 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274547 2 0.0747 0.912 0.016 0.984 0.000
#> SRR1274548 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.797 0.000 0.000 1.000
#> SRR1274550 3 0.4902 0.825 0.092 0.064 0.844
#> SRR1274551 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274552 3 0.5085 0.822 0.092 0.072 0.836
#> SRR1274553 3 0.4995 0.824 0.092 0.068 0.840
#> SRR1274554 2 0.3267 0.848 0.000 0.884 0.116
#> SRR1274555 3 0.3482 0.801 0.128 0.000 0.872
#> SRR1274556 3 0.2878 0.810 0.096 0.000 0.904
#> SRR1274557 3 0.3482 0.801 0.128 0.000 0.872
#> SRR1274558 2 0.3192 0.834 0.112 0.888 0.000
#> SRR1274559 1 0.2261 0.893 0.932 0.068 0.000
#> SRR1274560 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274561 1 0.5905 0.427 0.648 0.000 0.352
#> SRR1274562 1 0.5905 0.427 0.648 0.000 0.352
#> SRR1274563 3 0.3482 0.801 0.128 0.000 0.872
#> SRR1274564 2 0.6608 0.488 0.016 0.628 0.356
#> SRR1274565 2 0.2269 0.896 0.016 0.944 0.040
#> SRR1274566 2 0.3769 0.858 0.016 0.880 0.104
#> SRR1274567 3 0.6096 0.572 0.016 0.280 0.704
#> SRR1274568 2 0.2066 0.880 0.060 0.940 0.000
#> SRR1274569 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274570 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274571 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274572 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274573 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274575 2 0.3528 0.864 0.016 0.892 0.092
#> SRR1274576 2 0.0000 0.918 0.000 1.000 0.000
#> SRR1274577 2 0.1753 0.896 0.000 0.952 0.048
#> SRR1274578 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274579 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274580 1 0.1860 0.912 0.948 0.052 0.000
#> SRR1274581 2 0.7931 0.480 0.284 0.624 0.092
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.3710 0.6935 0.004 0.804 0.000 0.192
#> SRR1274530 2 0.3710 0.6935 0.004 0.804 0.000 0.192
#> SRR1274531 4 0.5243 0.3104 0.004 0.004 0.416 0.576
#> SRR1274532 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.6730 0.4733 0.008 0.280 0.104 0.608
#> SRR1274534 2 0.5368 0.4274 0.000 0.636 0.024 0.340
#> SRR1274535 4 0.5933 0.2199 0.000 0.408 0.040 0.552
#> SRR1274536 4 0.5935 0.4208 0.004 0.064 0.268 0.664
#> SRR1274537 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274539 2 0.5403 0.4111 0.000 0.628 0.024 0.348
#> SRR1274540 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.3105 0.7343 0.004 0.856 0.000 0.140
#> SRR1274546 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0376 0.8176 0.004 0.992 0.000 0.004
#> SRR1274548 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.4898 -0.0120 0.000 0.000 0.584 0.416
#> SRR1274550 4 0.5515 0.3891 0.008 0.008 0.420 0.564
#> SRR1274551 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274552 4 0.5702 0.4148 0.012 0.012 0.400 0.576
#> SRR1274553 4 0.5702 0.4148 0.012 0.012 0.400 0.576
#> SRR1274554 2 0.5663 0.1645 0.000 0.536 0.024 0.440
#> SRR1274555 3 0.0188 0.6703 0.000 0.000 0.996 0.004
#> SRR1274556 3 0.3219 0.4994 0.000 0.000 0.836 0.164
#> SRR1274557 3 0.0188 0.6703 0.000 0.000 0.996 0.004
#> SRR1274558 2 0.4395 0.6197 0.204 0.776 0.004 0.016
#> SRR1274559 1 0.2075 0.8954 0.936 0.016 0.004 0.044
#> SRR1274560 1 0.1247 0.9223 0.968 0.012 0.004 0.016
#> SRR1274561 3 0.4679 0.4438 0.352 0.000 0.648 0.000
#> SRR1274562 3 0.5291 0.4551 0.324 0.000 0.652 0.024
#> SRR1274563 3 0.0000 0.6700 0.000 0.000 1.000 0.000
#> SRR1274564 4 0.5873 0.4025 0.004 0.292 0.052 0.652
#> SRR1274565 2 0.4761 0.5607 0.004 0.664 0.000 0.332
#> SRR1274566 2 0.5119 0.3643 0.004 0.556 0.000 0.440
#> SRR1274567 4 0.5722 0.3889 0.004 0.044 0.292 0.660
#> SRR1274568 2 0.6429 0.4878 0.160 0.648 0.000 0.192
#> SRR1274569 1 0.0469 0.9311 0.988 0.012 0.000 0.000
#> SRR1274570 1 0.0469 0.9311 0.988 0.012 0.000 0.000
#> SRR1274571 1 0.0469 0.9311 0.988 0.012 0.000 0.000
#> SRR1274572 1 0.1767 0.9254 0.944 0.012 0.000 0.044
#> SRR1274573 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.3710 0.6935 0.004 0.804 0.000 0.192
#> SRR1274576 2 0.0000 0.8218 0.000 1.000 0.000 0.000
#> SRR1274577 2 0.5110 0.4268 0.000 0.636 0.012 0.352
#> SRR1274578 1 0.3529 0.8915 0.836 0.012 0.000 0.152
#> SRR1274579 1 0.3428 0.8962 0.844 0.012 0.000 0.144
#> SRR1274580 1 0.3428 0.8962 0.844 0.012 0.000 0.144
#> SRR1274581 2 0.8334 0.0798 0.152 0.424 0.044 0.380
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.4243 0.5172 0.000 0.712 0.000 0.264 0.024
#> SRR1274530 2 0.4243 0.5172 0.000 0.712 0.000 0.264 0.024
#> SRR1274531 4 0.5269 0.1136 0.004 0.000 0.120 0.688 0.188
#> SRR1274532 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.6473 0.4803 0.004 0.116 0.012 0.376 0.492
#> SRR1274534 2 0.6554 -0.0961 0.000 0.484 0.004 0.192 0.320
#> SRR1274535 5 0.6740 0.3317 0.000 0.212 0.004 0.380 0.404
#> SRR1274536 4 0.2172 0.4430 0.000 0.020 0.060 0.916 0.004
#> SRR1274537 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.6576 -0.1079 0.000 0.480 0.004 0.196 0.320
#> SRR1274540 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.2505 0.7251 0.000 0.888 0.000 0.092 0.020
#> SRR1274546 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0566 0.8075 0.000 0.984 0.000 0.004 0.012
#> SRR1274548 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 4 0.4268 0.1870 0.000 0.000 0.344 0.648 0.008
#> SRR1274550 5 0.6439 0.5384 0.000 0.000 0.192 0.332 0.476
#> SRR1274551 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 5 0.6282 0.5561 0.000 0.000 0.164 0.340 0.496
#> SRR1274553 5 0.6282 0.5561 0.000 0.000 0.164 0.340 0.496
#> SRR1274554 4 0.6939 -0.2471 0.000 0.328 0.004 0.368 0.300
#> SRR1274555 3 0.0798 0.8013 0.000 0.000 0.976 0.016 0.008
#> SRR1274556 3 0.3810 0.6206 0.000 0.000 0.788 0.176 0.036
#> SRR1274557 3 0.0798 0.8013 0.000 0.000 0.976 0.016 0.008
#> SRR1274558 2 0.5881 0.4113 0.248 0.652 0.008 0.048 0.044
#> SRR1274559 1 0.3631 0.7377 0.852 0.008 0.016 0.072 0.052
#> SRR1274560 1 0.3108 0.7613 0.880 0.004 0.016 0.048 0.052
#> SRR1274561 3 0.4818 0.6329 0.260 0.000 0.688 0.004 0.048
#> SRR1274562 3 0.4477 0.6529 0.252 0.000 0.708 0.000 0.040
#> SRR1274563 3 0.1082 0.8000 0.000 0.000 0.964 0.008 0.028
#> SRR1274564 4 0.2408 0.4615 0.000 0.092 0.016 0.892 0.000
#> SRR1274565 4 0.5238 -0.0111 0.000 0.472 0.000 0.484 0.044
#> SRR1274566 4 0.4354 0.3746 0.000 0.256 0.000 0.712 0.032
#> SRR1274567 4 0.3164 0.4561 0.000 0.020 0.084 0.868 0.028
#> SRR1274568 2 0.7836 0.0907 0.224 0.484 0.004 0.172 0.116
#> SRR1274569 1 0.0740 0.8102 0.980 0.004 0.000 0.008 0.008
#> SRR1274570 1 0.0162 0.8135 0.996 0.004 0.000 0.000 0.000
#> SRR1274571 1 0.0162 0.8135 0.996 0.004 0.000 0.000 0.000
#> SRR1274572 1 0.2674 0.7908 0.856 0.004 0.000 0.000 0.140
#> SRR1274573 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8169 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.4243 0.5172 0.000 0.712 0.000 0.264 0.024
#> SRR1274576 2 0.0798 0.8024 0.000 0.976 0.000 0.008 0.016
#> SRR1274577 2 0.6690 -0.1939 0.000 0.432 0.000 0.268 0.300
#> SRR1274578 1 0.4318 0.7322 0.644 0.004 0.000 0.004 0.348
#> SRR1274579 1 0.4302 0.7348 0.648 0.004 0.000 0.004 0.344
#> SRR1274580 1 0.4302 0.7348 0.648 0.004 0.000 0.004 0.344
#> SRR1274581 5 0.7323 -0.0410 0.076 0.392 0.004 0.100 0.428
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.5500 0.01256 0.000 0.564 0.004 0.332 0.016 0.084
#> SRR1274530 2 0.5500 0.01256 0.000 0.564 0.004 0.332 0.016 0.084
#> SRR1274531 5 0.5116 -0.19038 0.000 0.000 0.020 0.464 0.476 0.040
#> SRR1274532 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.2497 0.53748 0.000 0.032 0.000 0.032 0.896 0.040
#> SRR1274534 5 0.6229 -0.00698 0.000 0.376 0.008 0.032 0.472 0.112
#> SRR1274535 5 0.4831 0.48694 0.000 0.100 0.008 0.040 0.740 0.112
#> SRR1274536 4 0.2726 0.69268 0.000 0.008 0.008 0.848 0.136 0.000
#> SRR1274537 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.6229 -0.00698 0.000 0.376 0.008 0.032 0.472 0.112
#> SRR1274540 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.3063 0.60421 0.000 0.860 0.004 0.040 0.016 0.080
#> SRR1274546 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.1442 0.72205 0.000 0.944 0.004 0.000 0.012 0.040
#> SRR1274548 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 4 0.5411 0.39921 0.000 0.000 0.256 0.588 0.152 0.004
#> SRR1274550 5 0.1781 0.52761 0.000 0.000 0.060 0.008 0.924 0.008
#> SRR1274551 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.1909 0.52260 0.000 0.000 0.052 0.004 0.920 0.024
#> SRR1274553 5 0.1909 0.52260 0.000 0.000 0.052 0.004 0.920 0.024
#> SRR1274554 5 0.7425 0.27378 0.000 0.172 0.012 0.168 0.456 0.192
#> SRR1274555 3 0.0547 0.76738 0.000 0.000 0.980 0.000 0.020 0.000
#> SRR1274556 3 0.3554 0.63246 0.000 0.000 0.808 0.108 0.080 0.004
#> SRR1274557 3 0.0692 0.76652 0.000 0.000 0.976 0.004 0.020 0.000
#> SRR1274558 2 0.6313 -0.26810 0.312 0.464 0.000 0.024 0.000 0.200
#> SRR1274559 1 0.3790 0.56932 0.752 0.000 0.004 0.024 0.004 0.216
#> SRR1274560 1 0.3534 0.58342 0.772 0.000 0.004 0.024 0.000 0.200
#> SRR1274561 3 0.6172 0.46263 0.316 0.000 0.520 0.056 0.000 0.108
#> SRR1274562 3 0.5646 0.58355 0.248 0.000 0.612 0.048 0.000 0.092
#> SRR1274563 3 0.2772 0.76073 0.000 0.000 0.876 0.036 0.020 0.068
#> SRR1274564 4 0.2664 0.69532 0.000 0.016 0.000 0.848 0.136 0.000
#> SRR1274565 4 0.6752 -0.16016 0.000 0.296 0.008 0.472 0.052 0.172
#> SRR1274566 4 0.4580 0.62196 0.000 0.072 0.008 0.768 0.064 0.088
#> SRR1274567 4 0.3532 0.68987 0.000 0.004 0.016 0.828 0.092 0.060
#> SRR1274568 2 0.8410 -0.42763 0.248 0.320 0.004 0.096 0.080 0.252
#> SRR1274569 1 0.1524 0.67033 0.932 0.000 0.000 0.008 0.000 0.060
#> SRR1274570 1 0.0146 0.68855 0.996 0.000 0.004 0.000 0.000 0.000
#> SRR1274571 1 0.0508 0.68702 0.984 0.000 0.004 0.012 0.000 0.000
#> SRR1274572 1 0.3426 0.64897 0.764 0.000 0.004 0.012 0.000 0.220
#> SRR1274573 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.77725 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.5500 0.01256 0.000 0.564 0.004 0.332 0.016 0.084
#> SRR1274576 2 0.1633 0.68932 0.000 0.932 0.000 0.000 0.024 0.044
#> SRR1274577 5 0.7546 0.06027 0.000 0.264 0.008 0.140 0.392 0.196
#> SRR1274578 1 0.4097 0.54805 0.500 0.000 0.000 0.008 0.000 0.492
#> SRR1274579 1 0.3864 0.56652 0.520 0.000 0.000 0.000 0.000 0.480
#> SRR1274580 1 0.3864 0.56652 0.520 0.000 0.000 0.000 0.000 0.480
#> SRR1274581 6 0.7451 0.00000 0.040 0.340 0.012 0.060 0.120 0.428
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "skmeans"]
# you can also extract it by
# res = res_list["CV:skmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'skmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 3.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.849 0.881 0.955 0.5034 0.497 0.497
#> 3 3 0.882 0.894 0.954 0.3124 0.734 0.515
#> 4 4 0.721 0.652 0.851 0.1098 0.881 0.670
#> 5 5 0.733 0.713 0.834 0.0584 0.890 0.625
#> 6 6 0.792 0.741 0.853 0.0376 0.969 0.852
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 3
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.9521 0.000 1.000
#> SRR1274529 2 0.0000 0.9521 0.000 1.000
#> SRR1274530 2 0.0000 0.9521 0.000 1.000
#> SRR1274531 1 0.0000 0.9460 1.000 0.000
#> SRR1274532 2 0.0000 0.9521 0.000 1.000
#> SRR1274533 2 0.0938 0.9423 0.012 0.988
#> SRR1274534 2 0.0000 0.9521 0.000 1.000
#> SRR1274535 2 0.5059 0.8412 0.112 0.888
#> SRR1274536 2 0.8608 0.5822 0.284 0.716
#> SRR1274537 2 0.0000 0.9521 0.000 1.000
#> SRR1274538 2 0.0000 0.9521 0.000 1.000
#> SRR1274539 2 0.0000 0.9521 0.000 1.000
#> SRR1274540 2 0.0000 0.9521 0.000 1.000
#> SRR1274541 2 0.0000 0.9521 0.000 1.000
#> SRR1274542 2 0.0000 0.9521 0.000 1.000
#> SRR1274543 2 0.0000 0.9521 0.000 1.000
#> SRR1274544 2 0.0000 0.9521 0.000 1.000
#> SRR1274545 2 0.0000 0.9521 0.000 1.000
#> SRR1274546 2 0.0000 0.9521 0.000 1.000
#> SRR1274547 2 0.0000 0.9521 0.000 1.000
#> SRR1274548 2 0.0000 0.9521 0.000 1.000
#> SRR1274549 1 0.0000 0.9460 1.000 0.000
#> SRR1274550 1 0.0000 0.9460 1.000 0.000
#> SRR1274551 2 0.0000 0.9521 0.000 1.000
#> SRR1274552 1 0.4161 0.8691 0.916 0.084
#> SRR1274553 1 0.0000 0.9460 1.000 0.000
#> SRR1274554 2 0.9996 -0.0343 0.488 0.512
#> SRR1274555 1 0.0000 0.9460 1.000 0.000
#> SRR1274556 1 0.0000 0.9460 1.000 0.000
#> SRR1274557 1 0.0000 0.9460 1.000 0.000
#> SRR1274558 2 0.9580 0.3586 0.380 0.620
#> SRR1274559 1 0.0000 0.9460 1.000 0.000
#> SRR1274560 1 0.0000 0.9460 1.000 0.000
#> SRR1274561 1 0.0000 0.9460 1.000 0.000
#> SRR1274562 1 0.0000 0.9460 1.000 0.000
#> SRR1274563 1 0.0000 0.9460 1.000 0.000
#> SRR1274564 2 0.0000 0.9521 0.000 1.000
#> SRR1274565 2 0.0000 0.9521 0.000 1.000
#> SRR1274566 2 0.0000 0.9521 0.000 1.000
#> SRR1274567 1 0.0000 0.9460 1.000 0.000
#> SRR1274568 1 0.8608 0.5916 0.716 0.284
#> SRR1274569 1 0.0000 0.9460 1.000 0.000
#> SRR1274570 1 0.0000 0.9460 1.000 0.000
#> SRR1274571 1 0.0000 0.9460 1.000 0.000
#> SRR1274572 1 0.0000 0.9460 1.000 0.000
#> SRR1274573 2 0.0000 0.9521 0.000 1.000
#> SRR1274574 2 0.0000 0.9521 0.000 1.000
#> SRR1274575 2 0.0000 0.9521 0.000 1.000
#> SRR1274576 2 0.0000 0.9521 0.000 1.000
#> SRR1274577 1 0.9635 0.3756 0.612 0.388
#> SRR1274578 1 0.0000 0.9460 1.000 0.000
#> SRR1274579 1 0.0000 0.9460 1.000 0.000
#> SRR1274580 1 0.0000 0.9460 1.000 0.000
#> SRR1274581 1 0.9710 0.3197 0.600 0.400
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274529 2 0.1163 0.954 0.000 0.972 0.028
#> SRR1274530 2 0.1163 0.954 0.000 0.972 0.028
#> SRR1274531 3 0.0000 0.938 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274533 3 0.1163 0.931 0.000 0.028 0.972
#> SRR1274534 2 0.4974 0.685 0.000 0.764 0.236
#> SRR1274535 3 0.3038 0.859 0.000 0.104 0.896
#> SRR1274536 3 0.0000 0.938 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274539 2 0.5016 0.678 0.000 0.760 0.240
#> SRR1274540 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274545 2 0.1163 0.954 0.000 0.972 0.028
#> SRR1274546 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.938 0.000 0.000 1.000
#> SRR1274550 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274551 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274552 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274553 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274554 3 0.1163 0.931 0.000 0.028 0.972
#> SRR1274555 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274556 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274557 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274558 1 0.4931 0.698 0.768 0.232 0.000
#> SRR1274559 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274560 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274561 1 0.4504 0.730 0.804 0.000 0.196
#> SRR1274562 1 0.4504 0.730 0.804 0.000 0.196
#> SRR1274563 3 0.1163 0.940 0.028 0.000 0.972
#> SRR1274564 3 0.0592 0.932 0.000 0.012 0.988
#> SRR1274565 2 0.1163 0.954 0.000 0.972 0.028
#> SRR1274566 3 0.6225 0.219 0.000 0.432 0.568
#> SRR1274567 3 0.0000 0.938 0.000 0.000 1.000
#> SRR1274568 1 0.1163 0.896 0.972 0.028 0.000
#> SRR1274569 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274575 2 0.1163 0.954 0.000 0.972 0.028
#> SRR1274576 2 0.0000 0.969 0.000 1.000 0.000
#> SRR1274577 1 0.1163 0.896 0.972 0.028 0.000
#> SRR1274578 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.911 1.000 0.000 0.000
#> SRR1274581 1 0.6513 0.304 0.592 0.400 0.008
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.4643 0.5262 0.000 0.656 0.000 0.344
#> SRR1274530 2 0.4643 0.5262 0.000 0.656 0.000 0.344
#> SRR1274531 3 0.4998 0.0916 0.000 0.000 0.512 0.488
#> SRR1274532 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.0188 0.6071 0.000 0.004 0.996 0.000
#> SRR1274534 3 0.4343 0.4123 0.000 0.264 0.732 0.004
#> SRR1274535 3 0.1635 0.5893 0.000 0.044 0.948 0.008
#> SRR1274536 4 0.3219 0.6046 0.000 0.000 0.164 0.836
#> SRR1274537 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274539 3 0.4343 0.4123 0.000 0.264 0.732 0.004
#> SRR1274540 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.2814 0.8145 0.000 0.868 0.000 0.132
#> SRR1274546 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274549 4 0.4843 0.1141 0.000 0.000 0.396 0.604
#> SRR1274550 3 0.0000 0.6076 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.0000 0.6076 0.000 0.000 1.000 0.000
#> SRR1274553 3 0.0000 0.6076 0.000 0.000 1.000 0.000
#> SRR1274554 3 0.3975 0.4962 0.000 0.000 0.760 0.240
#> SRR1274555 3 0.4907 0.2823 0.000 0.000 0.580 0.420
#> SRR1274556 3 0.4907 0.2823 0.000 0.000 0.580 0.420
#> SRR1274557 3 0.4907 0.2823 0.000 0.000 0.580 0.420
#> SRR1274558 1 0.4331 0.5620 0.712 0.288 0.000 0.000
#> SRR1274559 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.5233 0.4324 0.648 0.000 0.020 0.332
#> SRR1274562 1 0.5331 0.4258 0.644 0.000 0.024 0.332
#> SRR1274563 3 0.4907 0.2823 0.000 0.000 0.580 0.420
#> SRR1274564 4 0.3123 0.6103 0.000 0.000 0.156 0.844
#> SRR1274565 4 0.4888 -0.0553 0.000 0.412 0.000 0.588
#> SRR1274566 4 0.0376 0.6230 0.000 0.004 0.004 0.992
#> SRR1274567 4 0.0336 0.6252 0.000 0.000 0.008 0.992
#> SRR1274568 1 0.2760 0.7433 0.872 0.128 0.000 0.000
#> SRR1274569 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.8419 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9309 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.4643 0.5262 0.000 0.656 0.000 0.344
#> SRR1274576 2 0.0188 0.9276 0.000 0.996 0.000 0.004
#> SRR1274577 3 0.7727 0.0122 0.388 0.004 0.416 0.192
#> SRR1274578 1 0.0336 0.8395 0.992 0.000 0.000 0.008
#> SRR1274579 1 0.0188 0.8411 0.996 0.000 0.000 0.004
#> SRR1274580 1 0.0188 0.8411 0.996 0.000 0.000 0.004
#> SRR1274581 1 0.8823 0.0931 0.404 0.368 0.140 0.088
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 4 0.4283 0.396 0.000 0.456 0.000 0.544 0.000
#> SRR1274530 4 0.4283 0.396 0.000 0.456 0.000 0.544 0.000
#> SRR1274531 3 0.1399 0.753 0.000 0.000 0.952 0.020 0.028
#> SRR1274532 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.3612 0.671 0.000 0.000 0.268 0.000 0.732
#> SRR1274534 5 0.4389 0.618 0.000 0.184 0.056 0.004 0.756
#> SRR1274535 5 0.3630 0.680 0.000 0.016 0.204 0.000 0.780
#> SRR1274536 4 0.4252 0.493 0.000 0.000 0.280 0.700 0.020
#> SRR1274537 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.4389 0.618 0.000 0.184 0.056 0.004 0.756
#> SRR1274540 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.3336 0.594 0.000 0.772 0.000 0.228 0.000
#> SRR1274546 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0404 0.960 0.000 0.988 0.000 0.012 0.000
#> SRR1274548 2 0.0162 0.967 0.000 0.996 0.000 0.000 0.004
#> SRR1274549 3 0.2798 0.682 0.000 0.000 0.852 0.140 0.008
#> SRR1274550 5 0.4235 0.567 0.000 0.000 0.424 0.000 0.576
#> SRR1274551 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 5 0.4182 0.598 0.000 0.000 0.400 0.000 0.600
#> SRR1274553 5 0.4182 0.598 0.000 0.000 0.400 0.000 0.600
#> SRR1274554 5 0.6071 0.525 0.000 0.000 0.192 0.236 0.572
#> SRR1274555 3 0.0000 0.782 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0000 0.782 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.0000 0.782 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 1 0.4236 0.560 0.728 0.248 0.000 0.008 0.016
#> SRR1274559 1 0.0798 0.781 0.976 0.000 0.000 0.008 0.016
#> SRR1274560 1 0.0798 0.781 0.976 0.000 0.000 0.008 0.016
#> SRR1274561 3 0.4359 0.393 0.412 0.000 0.584 0.000 0.004
#> SRR1274562 3 0.4310 0.422 0.392 0.000 0.604 0.000 0.004
#> SRR1274563 3 0.0000 0.782 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.4114 0.503 0.000 0.000 0.272 0.712 0.016
#> SRR1274565 4 0.3053 0.605 0.000 0.128 0.008 0.852 0.012
#> SRR1274566 4 0.3022 0.571 0.000 0.004 0.136 0.848 0.012
#> SRR1274567 4 0.3203 0.558 0.000 0.000 0.168 0.820 0.012
#> SRR1274568 1 0.4103 0.696 0.812 0.056 0.000 0.024 0.108
#> SRR1274569 1 0.0000 0.786 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.786 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.786 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0566 0.784 0.984 0.000 0.000 0.012 0.004
#> SRR1274573 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.971 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.4283 0.396 0.000 0.456 0.000 0.544 0.000
#> SRR1274576 2 0.2286 0.836 0.000 0.888 0.000 0.004 0.108
#> SRR1274577 5 0.5598 0.267 0.112 0.000 0.000 0.276 0.612
#> SRR1274578 1 0.5682 0.666 0.668 0.000 0.016 0.132 0.184
#> SRR1274579 1 0.5682 0.666 0.668 0.000 0.016 0.132 0.184
#> SRR1274580 1 0.5682 0.666 0.668 0.000 0.016 0.132 0.184
#> SRR1274581 1 0.8957 0.138 0.336 0.284 0.032 0.160 0.188
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.5390 0.340 0.000 0.424 0.012 0.504 0.016 0.044
#> SRR1274530 4 0.5390 0.340 0.000 0.424 0.012 0.504 0.016 0.044
#> SRR1274531 3 0.2333 0.814 0.000 0.000 0.896 0.060 0.040 0.004
#> SRR1274532 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.2350 0.792 0.000 0.000 0.100 0.000 0.880 0.020
#> SRR1274534 5 0.2492 0.746 0.000 0.068 0.000 0.008 0.888 0.036
#> SRR1274535 5 0.1465 0.777 0.000 0.004 0.024 0.004 0.948 0.020
#> SRR1274536 4 0.3176 0.511 0.000 0.000 0.156 0.812 0.032 0.000
#> SRR1274537 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0508 0.950 0.000 0.984 0.000 0.000 0.004 0.012
#> SRR1274539 5 0.2434 0.749 0.000 0.064 0.000 0.008 0.892 0.036
#> SRR1274540 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.4515 0.552 0.000 0.720 0.012 0.212 0.012 0.044
#> SRR1274546 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.2577 0.859 0.000 0.896 0.012 0.036 0.012 0.044
#> SRR1274548 2 0.0405 0.953 0.000 0.988 0.000 0.000 0.004 0.008
#> SRR1274549 3 0.3629 0.644 0.000 0.000 0.724 0.260 0.016 0.000
#> SRR1274550 5 0.3101 0.720 0.000 0.000 0.244 0.000 0.756 0.000
#> SRR1274551 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.3284 0.764 0.000 0.000 0.196 0.000 0.784 0.020
#> SRR1274553 5 0.3284 0.764 0.000 0.000 0.196 0.000 0.784 0.020
#> SRR1274554 5 0.6236 0.362 0.000 0.000 0.036 0.288 0.512 0.164
#> SRR1274555 3 0.0508 0.849 0.000 0.000 0.984 0.004 0.012 0.000
#> SRR1274556 3 0.0508 0.849 0.000 0.000 0.984 0.004 0.012 0.000
#> SRR1274557 3 0.0508 0.849 0.000 0.000 0.984 0.004 0.012 0.000
#> SRR1274558 1 0.1753 0.729 0.912 0.084 0.000 0.000 0.000 0.004
#> SRR1274559 1 0.0000 0.815 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.815 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274561 3 0.4127 0.575 0.284 0.000 0.684 0.004 0.000 0.028
#> SRR1274562 3 0.3857 0.709 0.160 0.000 0.772 0.004 0.000 0.064
#> SRR1274563 3 0.0363 0.847 0.000 0.000 0.988 0.000 0.012 0.000
#> SRR1274564 4 0.2911 0.524 0.000 0.000 0.144 0.832 0.024 0.000
#> SRR1274565 4 0.2752 0.533 0.000 0.012 0.000 0.864 0.020 0.104
#> SRR1274566 4 0.2508 0.547 0.000 0.000 0.016 0.884 0.016 0.084
#> SRR1274567 4 0.2831 0.544 0.000 0.000 0.032 0.868 0.016 0.084
#> SRR1274568 1 0.2357 0.772 0.888 0.012 0.000 0.004 0.004 0.092
#> SRR1274569 1 0.2558 0.811 0.840 0.000 0.000 0.004 0.000 0.156
#> SRR1274570 1 0.2632 0.807 0.832 0.000 0.000 0.004 0.000 0.164
#> SRR1274571 1 0.2773 0.805 0.828 0.000 0.004 0.004 0.000 0.164
#> SRR1274572 1 0.3756 0.593 0.676 0.000 0.004 0.004 0.000 0.316
#> SRR1274573 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.960 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 4 0.5390 0.340 0.000 0.424 0.012 0.504 0.016 0.044
#> SRR1274576 2 0.2742 0.823 0.000 0.872 0.000 0.008 0.076 0.044
#> SRR1274577 6 0.4559 0.406 0.008 0.000 0.004 0.092 0.172 0.724
#> SRR1274578 6 0.2964 0.666 0.204 0.000 0.004 0.000 0.000 0.792
#> SRR1274579 6 0.2964 0.666 0.204 0.000 0.004 0.000 0.000 0.792
#> SRR1274580 6 0.2964 0.666 0.204 0.000 0.004 0.000 0.000 0.792
#> SRR1274581 6 0.6972 0.333 0.084 0.280 0.008 0.076 0.032 0.520
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "pam"]
# you can also extract it by
# res = res_list["CV:pam"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'pam' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 6.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.958 0.982 0.3519 0.628 0.628
#> 3 3 0.968 0.928 0.978 0.3136 0.876 0.804
#> 4 4 0.966 0.927 0.977 0.0714 0.983 0.967
#> 5 5 0.942 0.902 0.969 0.2972 0.845 0.691
#> 6 6 0.905 0.913 0.958 0.0688 0.957 0.877
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 6
#> attr(,"optional")
#> [1] 2 3 5
There is also optional best \(k\) = 2 3 5 that is worth to check.
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 1.000 0.000 1.000
#> SRR1274529 2 0.000 1.000 0.000 1.000
#> SRR1274530 2 0.000 1.000 0.000 1.000
#> SRR1274531 2 0.000 1.000 0.000 1.000
#> SRR1274532 2 0.000 1.000 0.000 1.000
#> SRR1274533 2 0.000 1.000 0.000 1.000
#> SRR1274534 2 0.000 1.000 0.000 1.000
#> SRR1274535 2 0.000 1.000 0.000 1.000
#> SRR1274536 2 0.000 1.000 0.000 1.000
#> SRR1274537 2 0.000 1.000 0.000 1.000
#> SRR1274538 2 0.000 1.000 0.000 1.000
#> SRR1274539 2 0.000 1.000 0.000 1.000
#> SRR1274540 2 0.000 1.000 0.000 1.000
#> SRR1274541 2 0.000 1.000 0.000 1.000
#> SRR1274542 2 0.000 1.000 0.000 1.000
#> SRR1274543 2 0.000 1.000 0.000 1.000
#> SRR1274544 2 0.000 1.000 0.000 1.000
#> SRR1274545 2 0.000 1.000 0.000 1.000
#> SRR1274546 2 0.000 1.000 0.000 1.000
#> SRR1274547 2 0.000 1.000 0.000 1.000
#> SRR1274548 2 0.000 1.000 0.000 1.000
#> SRR1274549 2 0.000 1.000 0.000 1.000
#> SRR1274550 2 0.000 1.000 0.000 1.000
#> SRR1274551 2 0.000 1.000 0.000 1.000
#> SRR1274552 2 0.000 1.000 0.000 1.000
#> SRR1274553 2 0.000 1.000 0.000 1.000
#> SRR1274554 2 0.000 1.000 0.000 1.000
#> SRR1274555 1 0.975 0.398 0.592 0.408
#> SRR1274556 2 0.000 1.000 0.000 1.000
#> SRR1274557 2 0.000 1.000 0.000 1.000
#> SRR1274558 2 0.000 1.000 0.000 1.000
#> SRR1274559 1 0.697 0.760 0.812 0.188
#> SRR1274560 1 0.000 0.916 1.000 0.000
#> SRR1274561 1 0.000 0.916 1.000 0.000
#> SRR1274562 1 0.000 0.916 1.000 0.000
#> SRR1274563 1 0.000 0.916 1.000 0.000
#> SRR1274564 2 0.000 1.000 0.000 1.000
#> SRR1274565 2 0.000 1.000 0.000 1.000
#> SRR1274566 2 0.000 1.000 0.000 1.000
#> SRR1274567 2 0.000 1.000 0.000 1.000
#> SRR1274568 2 0.000 1.000 0.000 1.000
#> SRR1274569 1 0.000 0.916 1.000 0.000
#> SRR1274570 1 0.000 0.916 1.000 0.000
#> SRR1274571 1 0.000 0.916 1.000 0.000
#> SRR1274572 1 0.000 0.916 1.000 0.000
#> SRR1274573 2 0.000 1.000 0.000 1.000
#> SRR1274574 2 0.000 1.000 0.000 1.000
#> SRR1274575 2 0.000 1.000 0.000 1.000
#> SRR1274576 2 0.000 1.000 0.000 1.000
#> SRR1274577 2 0.000 1.000 0.000 1.000
#> SRR1274578 1 0.971 0.417 0.600 0.400
#> SRR1274579 1 0.000 0.916 1.000 0.000
#> SRR1274580 1 0.000 0.916 1.000 0.000
#> SRR1274581 2 0.000 1.000 0.000 1.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274529 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274530 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274531 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274532 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274533 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274534 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274535 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274536 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274537 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274538 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274539 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274540 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274541 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274542 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274543 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274544 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274545 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274546 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274547 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274548 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274549 3 0.406 0.685 0.000 0.164 0.836
#> SRR1274550 3 0.613 0.384 0.000 0.400 0.600
#> SRR1274551 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274552 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274553 2 0.129 0.963 0.000 0.968 0.032
#> SRR1274554 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274555 3 0.000 0.807 0.000 0.000 1.000
#> SRR1274556 3 0.000 0.807 0.000 0.000 1.000
#> SRR1274557 3 0.000 0.807 0.000 0.000 1.000
#> SRR1274558 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274559 1 0.450 0.619 0.804 0.196 0.000
#> SRR1274560 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274561 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274562 1 0.116 0.874 0.972 0.000 0.028
#> SRR1274563 3 0.000 0.807 0.000 0.000 1.000
#> SRR1274564 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274565 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274566 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274567 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274568 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274569 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274570 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274571 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274572 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274573 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274574 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274575 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274576 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274577 2 0.000 0.999 0.000 1.000 0.000
#> SRR1274578 1 0.610 0.264 0.608 0.392 0.000
#> SRR1274579 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274580 1 0.000 0.894 1.000 0.000 0.000
#> SRR1274581 2 0.000 0.999 0.000 1.000 0.000
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.0336 0.983 0.000 0.992 0.000 0.008
#> SRR1274531 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274532 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274533 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274534 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274535 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274536 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274537 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274539 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274540 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.3790 0.587 0.000 0.164 0.820 0.016
#> SRR1274550 3 0.4855 0.306 0.000 0.400 0.600 0.000
#> SRR1274551 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274552 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274553 2 0.1022 0.958 0.000 0.968 0.032 0.000
#> SRR1274554 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274555 3 0.0000 0.740 0.000 0.000 1.000 0.000
#> SRR1274556 3 0.0000 0.740 0.000 0.000 1.000 0.000
#> SRR1274557 3 0.0000 0.740 0.000 0.000 1.000 0.000
#> SRR1274558 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274559 1 0.3444 0.600 0.816 0.184 0.000 0.000
#> SRR1274560 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274562 1 0.0336 0.938 0.992 0.000 0.008 0.000
#> SRR1274563 3 0.0000 0.740 0.000 0.000 1.000 0.000
#> SRR1274564 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274565 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274566 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274567 2 0.0592 0.978 0.000 0.984 0.000 0.016
#> SRR1274568 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274569 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.944 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0336 0.983 0.000 0.992 0.000 0.008
#> SRR1274576 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274577 2 0.0000 0.987 0.000 1.000 0.000 0.000
#> SRR1274578 4 0.0592 1.000 0.016 0.000 0.000 0.984
#> SRR1274579 4 0.0592 1.000 0.016 0.000 0.000 0.984
#> SRR1274580 4 0.0592 1.000 0.016 0.000 0.000 0.984
#> SRR1274581 2 0.4331 0.582 0.000 0.712 0.000 0.288
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274530 2 0.0510 0.960 0.000 0.984 0.000 0.016 0.000
#> SRR1274531 4 0.0703 0.859 0.000 0.024 0.000 0.976 0.000
#> SRR1274532 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274534 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274535 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274536 4 0.0290 0.877 0.000 0.008 0.000 0.992 0.000
#> SRR1274537 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274540 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 4 0.0290 0.865 0.000 0.000 0.008 0.992 0.000
#> SRR1274550 3 0.4425 0.264 0.000 0.392 0.600 0.008 0.000
#> SRR1274551 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 2 0.0290 0.967 0.000 0.992 0.000 0.008 0.000
#> SRR1274553 2 0.1168 0.939 0.000 0.960 0.032 0.008 0.000
#> SRR1274554 4 0.4268 0.210 0.000 0.444 0.000 0.556 0.000
#> SRR1274555 3 0.0000 0.788 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0000 0.788 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.0000 0.788 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274559 1 0.2020 0.806 0.900 0.100 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274562 1 0.0290 0.966 0.992 0.000 0.008 0.000 0.000
#> SRR1274563 3 0.0000 0.788 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.0290 0.877 0.000 0.008 0.000 0.992 0.000
#> SRR1274565 4 0.0290 0.877 0.000 0.008 0.000 0.992 0.000
#> SRR1274566 4 0.0290 0.877 0.000 0.008 0.000 0.992 0.000
#> SRR1274567 4 0.0290 0.877 0.000 0.008 0.000 0.992 0.000
#> SRR1274568 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274569 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.972 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.1965 0.875 0.000 0.904 0.000 0.096 0.000
#> SRR1274576 2 0.0000 0.974 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 2 0.3210 0.704 0.000 0.788 0.000 0.212 0.000
#> SRR1274578 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274579 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274580 5 0.0000 1.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274581 2 0.3816 0.567 0.000 0.696 0.000 0.000 0.304
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274530 2 0.0458 0.948 0.000 0.984 0.000 0.016 0.000 0.000
#> SRR1274531 4 0.0458 0.864 0.000 0.016 0.000 0.984 0.000 0.000
#> SRR1274532 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 2 0.0790 0.934 0.000 0.968 0.000 0.000 0.032 0.000
#> SRR1274534 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274535 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274536 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274537 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274540 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274550 5 0.2340 1.000 0.000 0.148 0.000 0.000 0.852 0.000
#> SRR1274551 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.2340 1.000 0.000 0.148 0.000 0.000 0.852 0.000
#> SRR1274553 5 0.2340 1.000 0.000 0.148 0.000 0.000 0.852 0.000
#> SRR1274554 4 0.3823 0.140 0.000 0.436 0.000 0.564 0.000 0.000
#> SRR1274555 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274557 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274558 2 0.2340 0.794 0.000 0.852 0.000 0.000 0.148 0.000
#> SRR1274559 1 0.3967 0.754 0.760 0.092 0.000 0.000 0.148 0.000
#> SRR1274560 1 0.2340 0.874 0.852 0.000 0.000 0.000 0.148 0.000
#> SRR1274561 1 0.1814 0.900 0.900 0.000 0.000 0.000 0.100 0.000
#> SRR1274562 1 0.1007 0.912 0.956 0.000 0.044 0.000 0.000 0.000
#> SRR1274563 3 0.0000 1.000 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274565 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274566 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274567 4 0.0000 0.883 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274568 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274569 1 0.0000 0.927 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.927 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.927 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.927 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.1765 0.858 0.000 0.904 0.000 0.096 0.000 0.000
#> SRR1274576 2 0.0000 0.962 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 2 0.2969 0.659 0.000 0.776 0.000 0.224 0.000 0.000
#> SRR1274578 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274579 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274580 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274581 2 0.3565 0.497 0.000 0.692 0.000 0.004 0.000 0.304
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "mclust"]
# you can also extract it by
# res = res_list["CV:mclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'mclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.747 0.858 0.939 0.4397 0.575 0.575
#> 3 3 0.346 0.248 0.678 0.3211 0.640 0.476
#> 4 4 0.465 0.533 0.715 0.1654 0.748 0.531
#> 5 5 0.529 0.541 0.689 0.1093 0.646 0.258
#> 6 6 0.627 0.538 0.708 0.0601 0.843 0.416
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.933 0.000 1.000
#> SRR1274529 1 0.9795 0.337 0.584 0.416
#> SRR1274530 1 0.9833 0.315 0.576 0.424
#> SRR1274531 1 0.0672 0.927 0.992 0.008
#> SRR1274532 2 0.0000 0.933 0.000 1.000
#> SRR1274533 1 0.0672 0.927 0.992 0.008
#> SRR1274534 1 0.0000 0.931 1.000 0.000
#> SRR1274535 1 0.2778 0.896 0.952 0.048
#> SRR1274536 1 0.0000 0.931 1.000 0.000
#> SRR1274537 2 0.8327 0.682 0.264 0.736
#> SRR1274538 2 0.3733 0.898 0.072 0.928
#> SRR1274539 1 0.4298 0.860 0.912 0.088
#> SRR1274540 2 0.0000 0.933 0.000 1.000
#> SRR1274541 2 0.0000 0.933 0.000 1.000
#> SRR1274542 2 0.0000 0.933 0.000 1.000
#> SRR1274543 2 0.2948 0.911 0.052 0.948
#> SRR1274544 2 0.0000 0.933 0.000 1.000
#> SRR1274545 1 0.9996 0.105 0.512 0.488
#> SRR1274546 2 0.0000 0.933 0.000 1.000
#> SRR1274547 2 0.7745 0.693 0.228 0.772
#> SRR1274548 2 0.8144 0.701 0.252 0.748
#> SRR1274549 1 0.0000 0.931 1.000 0.000
#> SRR1274550 1 0.0672 0.927 0.992 0.008
#> SRR1274551 2 0.0376 0.933 0.004 0.996
#> SRR1274552 1 0.0672 0.927 0.992 0.008
#> SRR1274553 1 0.0376 0.929 0.996 0.004
#> SRR1274554 1 0.1184 0.922 0.984 0.016
#> SRR1274555 1 0.0000 0.931 1.000 0.000
#> SRR1274556 1 0.0000 0.931 1.000 0.000
#> SRR1274557 1 0.0000 0.931 1.000 0.000
#> SRR1274558 1 0.0000 0.931 1.000 0.000
#> SRR1274559 1 0.0000 0.931 1.000 0.000
#> SRR1274560 1 0.0000 0.931 1.000 0.000
#> SRR1274561 1 0.0000 0.931 1.000 0.000
#> SRR1274562 1 0.0000 0.931 1.000 0.000
#> SRR1274563 1 0.0000 0.931 1.000 0.000
#> SRR1274564 1 0.0000 0.931 1.000 0.000
#> SRR1274565 1 0.7950 0.680 0.760 0.240
#> SRR1274566 1 0.6623 0.770 0.828 0.172
#> SRR1274567 1 0.0000 0.931 1.000 0.000
#> SRR1274568 1 0.0000 0.931 1.000 0.000
#> SRR1274569 1 0.0000 0.931 1.000 0.000
#> SRR1274570 1 0.0000 0.931 1.000 0.000
#> SRR1274571 1 0.0000 0.931 1.000 0.000
#> SRR1274572 1 0.0000 0.931 1.000 0.000
#> SRR1274573 2 0.0000 0.933 0.000 1.000
#> SRR1274574 2 0.0376 0.933 0.004 0.996
#> SRR1274575 1 0.9795 0.337 0.584 0.416
#> SRR1274576 2 0.2778 0.913 0.048 0.952
#> SRR1274577 1 0.0000 0.931 1.000 0.000
#> SRR1274578 1 0.0000 0.931 1.000 0.000
#> SRR1274579 1 0.0000 0.931 1.000 0.000
#> SRR1274580 1 0.0000 0.931 1.000 0.000
#> SRR1274581 1 0.0376 0.929 0.996 0.004
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.362 0.67286 0.136 0.864 0.000
#> SRR1274529 2 0.586 0.52690 0.020 0.740 0.240
#> SRR1274530 2 0.536 0.55180 0.020 0.784 0.196
#> SRR1274531 3 0.620 -0.02130 0.000 0.424 0.576
#> SRR1274532 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274533 2 0.966 0.37165 0.244 0.460 0.296
#> SRR1274534 2 0.914 0.23456 0.144 0.448 0.408
#> SRR1274535 2 0.966 0.37165 0.244 0.460 0.296
#> SRR1274536 3 0.843 -0.04564 0.260 0.136 0.604
#> SRR1274537 2 0.594 0.62506 0.064 0.784 0.152
#> SRR1274538 2 0.511 0.61951 0.008 0.780 0.212
#> SRR1274539 2 0.966 0.37165 0.244 0.460 0.296
#> SRR1274540 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274541 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274542 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274543 2 0.429 0.65393 0.008 0.840 0.152
#> SRR1274544 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274545 2 0.495 0.57925 0.016 0.808 0.176
#> SRR1274546 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274547 2 0.362 0.63051 0.000 0.864 0.136
#> SRR1274548 2 0.506 0.61983 0.008 0.784 0.208
#> SRR1274549 3 0.819 -0.11308 0.292 0.104 0.604
#> SRR1274550 3 0.784 -0.14986 0.052 0.456 0.492
#> SRR1274551 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274552 3 0.792 -0.16459 0.056 0.460 0.484
#> SRR1274553 3 0.792 -0.16459 0.056 0.460 0.484
#> SRR1274554 2 0.966 0.37165 0.244 0.460 0.296
#> SRR1274555 3 0.826 -0.30086 0.436 0.076 0.488
#> SRR1274556 3 0.755 -0.01429 0.044 0.396 0.560
#> SRR1274557 3 0.755 -0.01429 0.044 0.396 0.560
#> SRR1274558 3 0.502 -0.24325 0.240 0.000 0.760
#> SRR1274559 3 0.502 -0.24325 0.240 0.000 0.760
#> SRR1274560 3 0.502 -0.24325 0.240 0.000 0.760
#> SRR1274561 1 0.620 0.94302 0.576 0.000 0.424
#> SRR1274562 1 0.620 0.94302 0.576 0.000 0.424
#> SRR1274563 1 0.604 0.88508 0.620 0.000 0.380
#> SRR1274564 3 0.613 0.01902 0.000 0.400 0.600
#> SRR1274565 2 0.652 0.15363 0.004 0.504 0.492
#> SRR1274566 3 0.621 -0.03045 0.000 0.428 0.572
#> SRR1274567 3 0.815 -0.12159 0.296 0.100 0.604
#> SRR1274568 3 0.502 -0.24325 0.240 0.000 0.760
#> SRR1274569 3 0.522 -0.23620 0.260 0.000 0.740
#> SRR1274570 3 0.525 -0.23969 0.264 0.000 0.736
#> SRR1274571 3 0.525 -0.23969 0.264 0.000 0.736
#> SRR1274572 3 0.525 -0.23969 0.264 0.000 0.736
#> SRR1274573 2 0.369 0.67297 0.140 0.860 0.000
#> SRR1274574 2 0.196 0.67880 0.056 0.944 0.000
#> SRR1274575 2 0.536 0.55180 0.020 0.784 0.196
#> SRR1274576 2 0.383 0.66282 0.008 0.868 0.124
#> SRR1274577 3 0.528 -0.24347 0.180 0.024 0.796
#> SRR1274578 3 0.514 -0.23790 0.252 0.000 0.748
#> SRR1274579 3 0.522 -0.23620 0.260 0.000 0.740
#> SRR1274580 3 0.522 -0.23620 0.260 0.000 0.740
#> SRR1274581 3 0.494 0.00975 0.056 0.104 0.840
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.4916 0.630 0.000 0.576 0.000 0.424
#> SRR1274529 2 0.5564 0.627 0.000 0.708 0.076 0.216
#> SRR1274530 2 0.5953 0.633 0.000 0.656 0.076 0.268
#> SRR1274531 2 0.5506 0.391 0.072 0.772 0.120 0.036
#> SRR1274532 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274533 2 0.6691 -0.522 0.016 0.560 0.060 0.364
#> SRR1274534 2 0.6529 0.192 0.056 0.696 0.068 0.180
#> SRR1274535 2 0.6495 -0.672 0.004 0.492 0.060 0.444
#> SRR1274536 2 0.5288 0.377 0.060 0.740 0.196 0.004
#> SRR1274537 2 0.3765 0.287 0.004 0.812 0.004 0.180
#> SRR1274538 2 0.0188 0.529 0.000 0.996 0.000 0.004
#> SRR1274539 2 0.6233 0.158 0.040 0.704 0.060 0.196
#> SRR1274540 2 0.4898 0.631 0.000 0.584 0.000 0.416
#> SRR1274541 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274542 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274543 2 0.1557 0.566 0.000 0.944 0.000 0.056
#> SRR1274544 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274545 2 0.5677 0.636 0.000 0.680 0.064 0.256
#> SRR1274546 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274547 2 0.4840 0.639 0.000 0.732 0.028 0.240
#> SRR1274548 2 0.0188 0.535 0.000 0.996 0.000 0.004
#> SRR1274549 2 0.6117 0.328 0.072 0.688 0.224 0.016
#> SRR1274550 4 0.7869 0.770 0.060 0.380 0.080 0.480
#> SRR1274551 2 0.4877 0.633 0.000 0.592 0.000 0.408
#> SRR1274552 4 0.7761 0.810 0.060 0.336 0.080 0.524
#> SRR1274553 4 0.7720 0.806 0.060 0.324 0.080 0.536
#> SRR1274554 2 0.6497 -0.678 0.004 0.488 0.060 0.448
#> SRR1274555 3 0.3668 0.670 0.072 0.024 0.872 0.032
#> SRR1274556 3 0.7392 0.239 0.072 0.364 0.524 0.040
#> SRR1274557 3 0.7172 0.385 0.072 0.304 0.584 0.040
#> SRR1274558 1 0.7696 0.621 0.536 0.028 0.300 0.136
#> SRR1274559 1 0.5254 0.746 0.672 0.028 0.300 0.000
#> SRR1274560 1 0.5231 0.750 0.676 0.028 0.296 0.000
#> SRR1274561 3 0.2760 0.625 0.128 0.000 0.872 0.000
#> SRR1274562 3 0.2760 0.625 0.128 0.000 0.872 0.000
#> SRR1274563 3 0.3497 0.630 0.124 0.000 0.852 0.024
#> SRR1274564 2 0.4606 0.439 0.060 0.808 0.124 0.008
#> SRR1274565 2 0.3304 0.476 0.048 0.888 0.052 0.012
#> SRR1274566 2 0.4147 0.451 0.040 0.832 0.120 0.008
#> SRR1274567 3 0.5325 0.515 0.068 0.204 0.728 0.000
#> SRR1274568 1 0.7153 0.698 0.604 0.028 0.264 0.104
#> SRR1274569 1 0.3266 0.833 0.832 0.000 0.168 0.000
#> SRR1274570 1 0.3266 0.833 0.832 0.000 0.168 0.000
#> SRR1274571 1 0.3266 0.833 0.832 0.000 0.168 0.000
#> SRR1274572 1 0.3266 0.833 0.832 0.000 0.168 0.000
#> SRR1274573 2 0.4925 0.630 0.000 0.572 0.000 0.428
#> SRR1274574 2 0.4605 0.639 0.000 0.664 0.000 0.336
#> SRR1274575 2 0.5873 0.632 0.000 0.668 0.076 0.256
#> SRR1274576 2 0.1389 0.558 0.000 0.952 0.000 0.048
#> SRR1274577 4 0.9028 0.632 0.292 0.256 0.064 0.388
#> SRR1274578 1 0.0707 0.765 0.980 0.000 0.020 0.000
#> SRR1274579 1 0.0000 0.762 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.762 1.000 0.000 0.000 0.000
#> SRR1274581 4 0.8623 0.752 0.168 0.376 0.056 0.400
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 4 0.0609 0.89698 0.000 0.020 0.000 0.980 0.000
#> SRR1274529 2 0.5120 0.32794 0.000 0.628 0.028 0.328 0.016
#> SRR1274530 2 0.4910 0.30404 0.000 0.628 0.020 0.340 0.012
#> SRR1274531 3 0.4596 0.55046 0.000 0.148 0.768 0.020 0.064
#> SRR1274532 4 0.0000 0.90802 0.000 0.000 0.000 1.000 0.000
#> SRR1274533 5 0.5615 0.88933 0.000 0.052 0.228 0.048 0.672
#> SRR1274534 5 0.7247 0.85555 0.036 0.088 0.208 0.076 0.592
#> SRR1274535 5 0.4665 0.87952 0.000 0.040 0.224 0.012 0.724
#> SRR1274536 3 0.4312 0.56229 0.000 0.156 0.776 0.008 0.060
#> SRR1274537 2 0.8014 0.43651 0.012 0.468 0.196 0.228 0.096
#> SRR1274538 2 0.7975 0.44443 0.012 0.404 0.204 0.316 0.064
#> SRR1274539 5 0.7097 0.86235 0.024 0.092 0.212 0.076 0.596
#> SRR1274540 4 0.0963 0.89193 0.000 0.036 0.000 0.964 0.000
#> SRR1274541 4 0.0000 0.90802 0.000 0.000 0.000 1.000 0.000
#> SRR1274542 4 0.0000 0.90802 0.000 0.000 0.000 1.000 0.000
#> SRR1274543 2 0.8174 0.42345 0.012 0.416 0.188 0.288 0.096
#> SRR1274544 4 0.0000 0.90802 0.000 0.000 0.000 1.000 0.000
#> SRR1274545 2 0.5136 0.21686 0.000 0.528 0.024 0.440 0.008
#> SRR1274546 4 0.0000 0.90802 0.000 0.000 0.000 1.000 0.000
#> SRR1274547 2 0.5774 0.24900 0.000 0.492 0.028 0.444 0.036
#> SRR1274548 2 0.7813 0.46424 0.012 0.464 0.200 0.260 0.064
#> SRR1274549 3 0.2625 0.60335 0.000 0.108 0.876 0.016 0.000
#> SRR1274550 3 0.4625 0.36299 0.000 0.012 0.660 0.012 0.316
#> SRR1274551 4 0.1341 0.87705 0.000 0.056 0.000 0.944 0.000
#> SRR1274552 3 0.4661 0.35242 0.000 0.012 0.652 0.012 0.324
#> SRR1274553 3 0.4643 0.35831 0.000 0.012 0.656 0.012 0.320
#> SRR1274554 5 0.4665 0.87952 0.000 0.040 0.224 0.012 0.724
#> SRR1274555 3 0.4210 0.53664 0.000 0.036 0.740 0.000 0.224
#> SRR1274556 3 0.1988 0.60822 0.000 0.048 0.928 0.008 0.016
#> SRR1274557 3 0.3071 0.60163 0.000 0.032 0.880 0.032 0.056
#> SRR1274558 2 0.6992 0.00298 0.224 0.564 0.136 0.000 0.076
#> SRR1274559 2 0.7057 -0.09343 0.256 0.532 0.156 0.000 0.056
#> SRR1274560 1 0.7188 0.46426 0.444 0.360 0.148 0.000 0.048
#> SRR1274561 3 0.5589 0.53595 0.036 0.076 0.684 0.000 0.204
#> SRR1274562 3 0.5308 0.53784 0.032 0.068 0.708 0.000 0.192
#> SRR1274563 3 0.4588 0.53819 0.004 0.056 0.732 0.000 0.208
#> SRR1274564 3 0.4650 0.54425 0.000 0.140 0.772 0.036 0.052
#> SRR1274565 2 0.8387 0.34182 0.052 0.408 0.328 0.140 0.072
#> SRR1274566 3 0.7513 -0.32038 0.000 0.364 0.420 0.128 0.088
#> SRR1274567 3 0.5300 0.57557 0.008 0.060 0.752 0.076 0.104
#> SRR1274568 2 0.6871 -0.02253 0.256 0.552 0.140 0.000 0.052
#> SRR1274569 1 0.3305 0.84686 0.776 0.224 0.000 0.000 0.000
#> SRR1274570 1 0.3461 0.84554 0.772 0.224 0.000 0.000 0.004
#> SRR1274571 1 0.3305 0.84686 0.776 0.224 0.000 0.000 0.000
#> SRR1274572 1 0.3305 0.84686 0.776 0.224 0.000 0.000 0.000
#> SRR1274573 4 0.0880 0.88091 0.000 0.032 0.000 0.968 0.000
#> SRR1274574 4 0.4822 -0.08201 0.000 0.416 0.016 0.564 0.004
#> SRR1274575 2 0.5042 0.32080 0.000 0.628 0.028 0.332 0.012
#> SRR1274576 2 0.8044 0.46115 0.012 0.440 0.196 0.268 0.084
#> SRR1274577 2 0.7968 -0.06788 0.076 0.328 0.312 0.000 0.284
#> SRR1274578 1 0.2127 0.73302 0.892 0.000 0.108 0.000 0.000
#> SRR1274579 1 0.0000 0.78515 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.78515 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 2 0.7348 -0.05751 0.024 0.352 0.332 0.000 0.292
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.1007 0.71099 0.000 0.956 0.000 0.000 0.000 0.044
#> SRR1274529 6 0.4178 0.43932 0.000 0.284 0.000 0.012 0.020 0.684
#> SRR1274530 6 0.4044 0.41906 0.000 0.312 0.000 0.008 0.012 0.668
#> SRR1274531 4 0.7439 0.61852 0.000 0.028 0.156 0.448 0.268 0.100
#> SRR1274532 2 0.0000 0.73254 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.2405 0.82010 0.000 0.016 0.004 0.000 0.880 0.100
#> SRR1274534 5 0.3351 0.78088 0.000 0.036 0.000 0.004 0.808 0.152
#> SRR1274535 5 0.1411 0.83263 0.000 0.004 0.000 0.000 0.936 0.060
#> SRR1274536 4 0.5968 0.66864 0.000 0.000 0.044 0.580 0.236 0.140
#> SRR1274537 6 0.5137 0.12012 0.000 0.416 0.000 0.004 0.072 0.508
#> SRR1274538 2 0.6108 -0.06095 0.000 0.424 0.000 0.004 0.240 0.332
#> SRR1274539 5 0.3313 0.77990 0.000 0.036 0.000 0.004 0.812 0.148
#> SRR1274540 2 0.0458 0.72528 0.000 0.984 0.000 0.000 0.000 0.016
#> SRR1274541 2 0.0000 0.73254 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.73254 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.6134 -0.06409 0.000 0.408 0.000 0.004 0.244 0.344
#> SRR1274544 2 0.0000 0.73254 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 6 0.4877 0.30691 0.000 0.424 0.000 0.036 0.012 0.528
#> SRR1274546 2 0.0146 0.73055 0.000 0.996 0.000 0.000 0.000 0.004
#> SRR1274547 6 0.4935 0.19745 0.000 0.460 0.000 0.004 0.052 0.484
#> SRR1274548 6 0.5160 0.06354 0.000 0.448 0.000 0.004 0.072 0.476
#> SRR1274549 4 0.5804 0.43505 0.004 0.000 0.332 0.532 0.116 0.016
#> SRR1274550 5 0.1801 0.80648 0.000 0.004 0.040 0.012 0.932 0.012
#> SRR1274551 2 0.1204 0.69346 0.000 0.944 0.000 0.000 0.000 0.056
#> SRR1274552 5 0.1672 0.81174 0.000 0.004 0.028 0.012 0.940 0.016
#> SRR1274553 5 0.1749 0.81006 0.000 0.004 0.032 0.012 0.936 0.016
#> SRR1274554 5 0.1333 0.83370 0.000 0.008 0.000 0.000 0.944 0.048
#> SRR1274555 3 0.0146 0.87922 0.000 0.004 0.996 0.000 0.000 0.000
#> SRR1274556 3 0.2256 0.86877 0.000 0.008 0.892 0.004 0.092 0.004
#> SRR1274557 3 0.2205 0.87353 0.000 0.008 0.896 0.004 0.088 0.004
#> SRR1274558 6 0.7558 -0.23732 0.316 0.000 0.032 0.268 0.056 0.328
#> SRR1274559 1 0.7209 -0.00272 0.412 0.000 0.040 0.288 0.028 0.232
#> SRR1274560 1 0.6593 0.17151 0.524 0.000 0.040 0.268 0.020 0.148
#> SRR1274561 4 0.5532 0.40407 0.096 0.000 0.304 0.580 0.004 0.016
#> SRR1274562 4 0.5447 0.35666 0.104 0.000 0.340 0.548 0.004 0.004
#> SRR1274563 3 0.0260 0.87412 0.008 0.000 0.992 0.000 0.000 0.000
#> SRR1274564 4 0.6798 0.65172 0.000 0.032 0.040 0.524 0.236 0.168
#> SRR1274565 4 0.7790 0.25952 0.016 0.136 0.000 0.332 0.216 0.300
#> SRR1274566 4 0.7316 0.60989 0.000 0.116 0.036 0.492 0.240 0.116
#> SRR1274567 4 0.6049 0.67589 0.008 0.008 0.072 0.612 0.236 0.064
#> SRR1274568 6 0.7561 -0.24787 0.268 0.000 0.036 0.268 0.056 0.372
#> SRR1274569 1 0.0146 0.70681 0.996 0.000 0.000 0.000 0.000 0.004
#> SRR1274570 1 0.0000 0.70719 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0146 0.70681 0.996 0.000 0.000 0.000 0.000 0.004
#> SRR1274572 1 0.0000 0.70719 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0146 0.73053 0.000 0.996 0.000 0.000 0.000 0.004
#> SRR1274574 2 0.3151 0.42814 0.000 0.748 0.000 0.000 0.000 0.252
#> SRR1274575 6 0.4097 0.43987 0.000 0.284 0.000 0.012 0.016 0.688
#> SRR1274576 2 0.6138 -0.06898 0.000 0.428 0.000 0.004 0.272 0.296
#> SRR1274577 5 0.4797 0.58555 0.016 0.000 0.008 0.032 0.644 0.300
#> SRR1274578 1 0.4264 0.61976 0.604 0.000 0.000 0.376 0.008 0.012
#> SRR1274579 1 0.4161 0.62056 0.608 0.000 0.000 0.376 0.004 0.012
#> SRR1274580 1 0.4161 0.62056 0.608 0.000 0.000 0.376 0.004 0.012
#> SRR1274581 5 0.4392 0.68060 0.012 0.004 0.000 0.044 0.716 0.224
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["CV", "NMF"]
# you can also extract it by
# res = res_list["CV:NMF"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'CV' method.
#> Subgroups are detected by 'NMF' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 3.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.641 0.847 0.926 0.4801 0.508 0.508
#> 3 3 0.875 0.876 0.950 0.3501 0.735 0.527
#> 4 4 0.632 0.455 0.716 0.1236 0.881 0.682
#> 5 5 0.760 0.725 0.826 0.0718 0.796 0.412
#> 6 6 0.880 0.828 0.916 0.0468 0.952 0.780
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 3
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.939 0.000 1.000
#> SRR1274529 2 0.0000 0.939 0.000 1.000
#> SRR1274530 2 0.0000 0.939 0.000 1.000
#> SRR1274531 1 0.9896 0.363 0.560 0.440
#> SRR1274532 2 0.0000 0.939 0.000 1.000
#> SRR1274533 2 0.0000 0.939 0.000 1.000
#> SRR1274534 2 0.0000 0.939 0.000 1.000
#> SRR1274535 2 0.0376 0.935 0.004 0.996
#> SRR1274536 2 0.0000 0.939 0.000 1.000
#> SRR1274537 2 0.3879 0.863 0.076 0.924
#> SRR1274538 2 0.0000 0.939 0.000 1.000
#> SRR1274539 2 0.0000 0.939 0.000 1.000
#> SRR1274540 2 0.0000 0.939 0.000 1.000
#> SRR1274541 2 0.0000 0.939 0.000 1.000
#> SRR1274542 2 0.0000 0.939 0.000 1.000
#> SRR1274543 2 0.0000 0.939 0.000 1.000
#> SRR1274544 2 0.0000 0.939 0.000 1.000
#> SRR1274545 2 0.0000 0.939 0.000 1.000
#> SRR1274546 2 0.0000 0.939 0.000 1.000
#> SRR1274547 2 0.0000 0.939 0.000 1.000
#> SRR1274548 2 0.0000 0.939 0.000 1.000
#> SRR1274549 1 0.7219 0.828 0.800 0.200
#> SRR1274550 1 0.7219 0.828 0.800 0.200
#> SRR1274551 2 0.0000 0.939 0.000 1.000
#> SRR1274552 1 0.7219 0.828 0.800 0.200
#> SRR1274553 1 0.7219 0.828 0.800 0.200
#> SRR1274554 2 0.9635 0.235 0.388 0.612
#> SRR1274555 1 0.7219 0.828 0.800 0.200
#> SRR1274556 1 0.7219 0.828 0.800 0.200
#> SRR1274557 1 0.7219 0.828 0.800 0.200
#> SRR1274558 2 0.9460 0.473 0.364 0.636
#> SRR1274559 1 0.0000 0.878 1.000 0.000
#> SRR1274560 1 0.0000 0.878 1.000 0.000
#> SRR1274561 1 0.0000 0.878 1.000 0.000
#> SRR1274562 1 0.0000 0.878 1.000 0.000
#> SRR1274563 1 0.7219 0.828 0.800 0.200
#> SRR1274564 2 0.0000 0.939 0.000 1.000
#> SRR1274565 2 0.0000 0.939 0.000 1.000
#> SRR1274566 2 0.0000 0.939 0.000 1.000
#> SRR1274567 1 0.7602 0.805 0.780 0.220
#> SRR1274568 2 0.9754 0.376 0.408 0.592
#> SRR1274569 1 0.0000 0.878 1.000 0.000
#> SRR1274570 1 0.0000 0.878 1.000 0.000
#> SRR1274571 1 0.0000 0.878 1.000 0.000
#> SRR1274572 1 0.0000 0.878 1.000 0.000
#> SRR1274573 2 0.0000 0.939 0.000 1.000
#> SRR1274574 2 0.0000 0.939 0.000 1.000
#> SRR1274575 2 0.0000 0.939 0.000 1.000
#> SRR1274576 2 0.0000 0.939 0.000 1.000
#> SRR1274577 2 0.9998 0.127 0.492 0.508
#> SRR1274578 1 0.0000 0.878 1.000 0.000
#> SRR1274579 1 0.0000 0.878 1.000 0.000
#> SRR1274580 1 0.0000 0.878 1.000 0.000
#> SRR1274581 1 0.0000 0.878 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274531 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274532 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274533 3 0.2878 0.880 0.000 0.096 0.904
#> SRR1274534 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274535 3 0.1529 0.938 0.000 0.040 0.960
#> SRR1274536 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274537 2 0.0237 0.948 0.004 0.996 0.000
#> SRR1274538 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274539 2 0.0237 0.949 0.000 0.996 0.004
#> SRR1274540 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274549 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274550 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274551 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274552 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274553 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274554 3 0.0424 0.964 0.000 0.008 0.992
#> SRR1274555 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274556 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274557 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274558 1 0.5948 0.439 0.640 0.360 0.000
#> SRR1274559 1 0.0000 0.893 1.000 0.000 0.000
#> SRR1274560 1 0.0000 0.893 1.000 0.000 0.000
#> SRR1274561 1 0.4796 0.696 0.780 0.000 0.220
#> SRR1274562 1 0.3941 0.778 0.844 0.000 0.156
#> SRR1274563 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274564 3 0.4931 0.682 0.000 0.232 0.768
#> SRR1274565 2 0.3879 0.796 0.000 0.848 0.152
#> SRR1274566 2 0.6291 0.120 0.000 0.532 0.468
#> SRR1274567 3 0.0237 0.967 0.000 0.004 0.996
#> SRR1274568 1 0.6180 0.298 0.584 0.416 0.000
#> SRR1274569 1 0.0000 0.893 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.893 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.893 1.000 0.000 0.000
#> SRR1274572 1 0.0237 0.893 0.996 0.000 0.004
#> SRR1274573 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.952 0.000 1.000 0.000
#> SRR1274577 2 0.6359 0.233 0.404 0.592 0.004
#> SRR1274578 1 0.0237 0.893 0.996 0.000 0.004
#> SRR1274579 1 0.0237 0.893 0.996 0.000 0.004
#> SRR1274580 1 0.0237 0.893 0.996 0.000 0.004
#> SRR1274581 1 0.2165 0.860 0.936 0.000 0.064
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.4804 0.53253 0.384 0.616 0.000 0.000
#> SRR1274530 2 0.4804 0.53253 0.384 0.616 0.000 0.000
#> SRR1274531 3 0.4103 0.62223 0.000 0.000 0.744 0.256
#> SRR1274532 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.2760 0.55577 0.000 0.128 0.872 0.000
#> SRR1274534 2 0.4933 0.32623 0.000 0.568 0.432 0.000
#> SRR1274535 3 0.3400 0.50643 0.000 0.180 0.820 0.000
#> SRR1274536 4 0.7514 -0.02030 0.384 0.000 0.184 0.432
#> SRR1274537 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0188 0.81793 0.000 0.996 0.004 0.000
#> SRR1274539 2 0.4933 0.32623 0.000 0.568 0.432 0.000
#> SRR1274540 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.4790 0.53611 0.380 0.620 0.000 0.000
#> SRR1274546 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.4331 0.61362 0.288 0.712 0.000 0.000
#> SRR1274548 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.5105 0.58971 0.004 0.000 0.564 0.432
#> SRR1274550 3 0.0000 0.62739 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.0188 0.62588 0.000 0.004 0.996 0.000
#> SRR1274553 3 0.0000 0.62739 0.000 0.000 1.000 0.000
#> SRR1274554 3 0.6152 0.57850 0.000 0.120 0.668 0.212
#> SRR1274555 3 0.4933 0.59386 0.000 0.000 0.568 0.432
#> SRR1274556 3 0.4933 0.59386 0.000 0.000 0.568 0.432
#> SRR1274557 3 0.4933 0.59386 0.000 0.000 0.568 0.432
#> SRR1274558 1 0.6489 0.31984 0.548 0.372 0.000 0.080
#> SRR1274559 1 0.4804 0.47845 0.616 0.000 0.000 0.384
#> SRR1274560 1 0.4804 0.47845 0.616 0.000 0.000 0.384
#> SRR1274561 4 0.4123 0.13465 0.220 0.000 0.008 0.772
#> SRR1274562 4 0.5594 0.22030 0.100 0.000 0.180 0.720
#> SRR1274563 3 0.4933 0.59386 0.000 0.000 0.568 0.432
#> SRR1274564 4 0.7514 -0.02030 0.384 0.000 0.184 0.432
#> SRR1274565 1 0.9223 -0.24505 0.384 0.304 0.088 0.224
#> SRR1274566 4 0.8156 -0.00363 0.384 0.032 0.156 0.428
#> SRR1274567 4 0.7514 -0.02030 0.384 0.000 0.184 0.432
#> SRR1274568 1 0.6407 0.29838 0.520 0.412 0.000 0.068
#> SRR1274569 1 0.4804 0.47845 0.616 0.000 0.000 0.384
#> SRR1274570 1 0.4804 0.47845 0.616 0.000 0.000 0.384
#> SRR1274571 1 0.4804 0.47845 0.616 0.000 0.000 0.384
#> SRR1274572 4 0.4985 -0.37758 0.468 0.000 0.000 0.532
#> SRR1274573 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0469 0.81414 0.012 0.988 0.000 0.000
#> SRR1274575 2 0.4804 0.53253 0.384 0.616 0.000 0.000
#> SRR1274576 2 0.0000 0.82024 0.000 1.000 0.000 0.000
#> SRR1274577 2 0.9663 -0.12073 0.156 0.360 0.276 0.208
#> SRR1274578 4 0.4933 -0.33476 0.432 0.000 0.000 0.568
#> SRR1274579 4 0.4933 -0.33476 0.432 0.000 0.000 0.568
#> SRR1274580 4 0.4933 -0.33476 0.432 0.000 0.000 0.568
#> SRR1274581 4 0.3464 0.21181 0.076 0.000 0.056 0.868
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 4 0.3730 0.6991 0.000 0.288 0.000 0.712 0.000
#> SRR1274530 4 0.3857 0.6806 0.000 0.312 0.000 0.688 0.000
#> SRR1274531 5 0.4452 0.3106 0.000 0.000 0.496 0.004 0.500
#> SRR1274532 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.3573 0.8353 0.000 0.036 0.152 0.000 0.812
#> SRR1274534 5 0.3003 0.7305 0.000 0.188 0.000 0.000 0.812
#> SRR1274535 5 0.3667 0.8317 0.000 0.048 0.140 0.000 0.812
#> SRR1274536 4 0.3730 0.6832 0.000 0.000 0.288 0.712 0.000
#> SRR1274537 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0404 0.8971 0.000 0.988 0.000 0.000 0.012
#> SRR1274539 5 0.3003 0.7305 0.000 0.188 0.000 0.000 0.812
#> SRR1274540 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 4 0.3895 0.6693 0.000 0.320 0.000 0.680 0.000
#> SRR1274546 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.3684 0.4436 0.000 0.720 0.000 0.280 0.000
#> SRR1274548 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 3 0.0880 0.8508 0.000 0.000 0.968 0.032 0.000
#> SRR1274550 5 0.3003 0.8298 0.000 0.000 0.188 0.000 0.812
#> SRR1274551 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 5 0.3003 0.8298 0.000 0.000 0.188 0.000 0.812
#> SRR1274553 5 0.3003 0.8298 0.000 0.000 0.188 0.000 0.812
#> SRR1274554 3 0.2149 0.8001 0.000 0.048 0.916 0.000 0.036
#> SRR1274555 3 0.0000 0.8656 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0000 0.8656 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.0000 0.8656 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 1 0.4171 0.1924 0.604 0.396 0.000 0.000 0.000
#> SRR1274559 1 0.0000 0.7347 1.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.7347 1.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.3752 0.3861 0.708 0.000 0.292 0.000 0.000
#> SRR1274562 3 0.3006 0.7484 0.156 0.000 0.836 0.004 0.004
#> SRR1274563 3 0.0000 0.8656 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.3730 0.6832 0.000 0.000 0.288 0.712 0.000
#> SRR1274565 4 0.5199 0.6556 0.000 0.072 0.292 0.636 0.000
#> SRR1274566 4 0.3730 0.6832 0.000 0.000 0.288 0.712 0.000
#> SRR1274567 4 0.3774 0.6775 0.000 0.000 0.296 0.704 0.000
#> SRR1274568 2 0.4283 0.1339 0.456 0.544 0.000 0.000 0.000
#> SRR1274569 1 0.0000 0.7347 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.7347 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.7347 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.4948 0.6545 0.676 0.000 0.000 0.256 0.068
#> SRR1274573 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0162 0.9044 0.000 0.996 0.000 0.004 0.000
#> SRR1274575 4 0.3837 0.6850 0.000 0.308 0.000 0.692 0.000
#> SRR1274576 2 0.0000 0.9080 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 2 0.7492 0.0744 0.064 0.460 0.000 0.288 0.188
#> SRR1274578 1 0.6284 0.5943 0.524 0.000 0.000 0.288 0.188
#> SRR1274579 1 0.6284 0.5943 0.524 0.000 0.000 0.288 0.188
#> SRR1274580 1 0.6284 0.5943 0.524 0.000 0.000 0.288 0.188
#> SRR1274581 3 0.7505 0.2773 0.108 0.000 0.484 0.284 0.124
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.2020 0.8297 0.000 0.096 0.000 0.896 0.000 0.008
#> SRR1274530 4 0.3298 0.7465 0.000 0.236 0.000 0.756 0.000 0.008
#> SRR1274531 3 0.4169 0.2104 0.000 0.000 0.532 0.000 0.456 0.012
#> SRR1274532 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0260 0.9834 0.000 0.000 0.000 0.000 0.992 0.008
#> SRR1274534 5 0.0891 0.9644 0.000 0.024 0.000 0.000 0.968 0.008
#> SRR1274535 5 0.0146 0.9828 0.000 0.000 0.000 0.000 0.996 0.004
#> SRR1274536 4 0.2020 0.8142 0.000 0.000 0.096 0.896 0.000 0.008
#> SRR1274537 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0260 0.9380 0.000 0.992 0.000 0.000 0.008 0.000
#> SRR1274539 5 0.0806 0.9692 0.000 0.020 0.000 0.000 0.972 0.008
#> SRR1274540 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0146 0.9411 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274544 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 4 0.3323 0.7421 0.000 0.240 0.000 0.752 0.000 0.008
#> SRR1274546 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.3103 0.6584 0.000 0.784 0.000 0.208 0.000 0.008
#> SRR1274548 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 3 0.3465 0.7426 0.000 0.000 0.804 0.156 0.016 0.024
#> SRR1274550 5 0.0146 0.9828 0.000 0.000 0.000 0.000 0.996 0.004
#> SRR1274551 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.0260 0.9834 0.000 0.000 0.000 0.000 0.992 0.008
#> SRR1274553 5 0.0260 0.9834 0.000 0.000 0.000 0.000 0.992 0.008
#> SRR1274554 3 0.2526 0.7594 0.000 0.004 0.876 0.096 0.000 0.024
#> SRR1274555 3 0.0806 0.8422 0.000 0.000 0.972 0.000 0.020 0.008
#> SRR1274556 3 0.0909 0.8407 0.000 0.000 0.968 0.000 0.020 0.012
#> SRR1274557 3 0.0458 0.8421 0.000 0.000 0.984 0.000 0.016 0.000
#> SRR1274558 1 0.1663 0.7962 0.912 0.088 0.000 0.000 0.000 0.000
#> SRR1274559 1 0.0000 0.8898 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.8898 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.1738 0.8417 0.928 0.000 0.052 0.004 0.000 0.016
#> SRR1274562 3 0.3862 0.7199 0.108 0.000 0.792 0.000 0.012 0.088
#> SRR1274563 3 0.0717 0.8423 0.000 0.000 0.976 0.000 0.016 0.008
#> SRR1274564 4 0.2020 0.8142 0.000 0.000 0.096 0.896 0.000 0.008
#> SRR1274565 4 0.2703 0.7677 0.000 0.028 0.080 0.876 0.000 0.016
#> SRR1274566 4 0.0547 0.8068 0.000 0.000 0.020 0.980 0.000 0.000
#> SRR1274567 4 0.1863 0.8117 0.000 0.000 0.104 0.896 0.000 0.000
#> SRR1274568 2 0.5418 0.1018 0.416 0.504 0.008 0.060 0.000 0.012
#> SRR1274569 1 0.0000 0.8898 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.8898 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.8898 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.3864 -0.0504 0.520 0.000 0.000 0.000 0.000 0.480
#> SRR1274573 2 0.0000 0.9436 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0146 0.9408 0.000 0.996 0.000 0.000 0.000 0.004
#> SRR1274575 4 0.2743 0.8030 0.000 0.164 0.000 0.828 0.000 0.008
#> SRR1274576 2 0.1732 0.8704 0.000 0.920 0.004 0.072 0.000 0.004
#> SRR1274577 6 0.3351 0.8008 0.004 0.040 0.020 0.096 0.000 0.840
#> SRR1274578 6 0.1075 0.8764 0.048 0.000 0.000 0.000 0.000 0.952
#> SRR1274579 6 0.1267 0.8763 0.060 0.000 0.000 0.000 0.000 0.940
#> SRR1274580 6 0.1387 0.8726 0.068 0.000 0.000 0.000 0.000 0.932
#> SRR1274581 6 0.5279 0.6539 0.100 0.000 0.168 0.052 0.000 0.680
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "hclust"]
# you can also extract it by
# res = res_list["MAD:hclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'hclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.902 0.958 0.975 0.4372 0.560 0.560
#> 3 3 0.923 0.932 0.980 0.0447 0.989 0.980
#> 4 4 0.812 0.898 0.929 0.1296 0.956 0.920
#> 5 5 0.661 0.771 0.909 0.0968 0.922 0.847
#> 6 6 0.721 0.819 0.915 0.1253 0.949 0.883
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.977 0.000 1.000
#> SRR1274529 2 0.000 0.977 0.000 1.000
#> SRR1274530 2 0.000 0.977 0.000 1.000
#> SRR1274531 2 0.000 0.977 0.000 1.000
#> SRR1274532 2 0.000 0.977 0.000 1.000
#> SRR1274533 2 0.000 0.977 0.000 1.000
#> SRR1274534 2 0.000 0.977 0.000 1.000
#> SRR1274535 2 0.000 0.977 0.000 1.000
#> SRR1274536 2 0.518 0.879 0.116 0.884
#> SRR1274537 2 0.000 0.977 0.000 1.000
#> SRR1274538 2 0.000 0.977 0.000 1.000
#> SRR1274539 2 0.000 0.977 0.000 1.000
#> SRR1274540 2 0.000 0.977 0.000 1.000
#> SRR1274541 2 0.000 0.977 0.000 1.000
#> SRR1274542 2 0.000 0.977 0.000 1.000
#> SRR1274543 2 0.000 0.977 0.000 1.000
#> SRR1274544 2 0.000 0.977 0.000 1.000
#> SRR1274545 2 0.000 0.977 0.000 1.000
#> SRR1274546 2 0.000 0.977 0.000 1.000
#> SRR1274547 2 0.000 0.977 0.000 1.000
#> SRR1274548 2 0.000 0.977 0.000 1.000
#> SRR1274549 2 0.795 0.705 0.240 0.760
#> SRR1274550 1 0.141 0.982 0.980 0.020
#> SRR1274551 2 0.000 0.977 0.000 1.000
#> SRR1274552 1 0.141 0.982 0.980 0.020
#> SRR1274553 1 0.141 0.982 0.980 0.020
#> SRR1274554 2 0.000 0.977 0.000 1.000
#> SRR1274555 1 0.141 0.982 0.980 0.020
#> SRR1274556 1 0.141 0.982 0.980 0.020
#> SRR1274557 1 0.141 0.982 0.980 0.020
#> SRR1274558 2 0.000 0.977 0.000 1.000
#> SRR1274559 2 0.373 0.923 0.072 0.928
#> SRR1274560 2 0.373 0.923 0.072 0.928
#> SRR1274561 1 0.844 0.636 0.728 0.272
#> SRR1274562 1 0.141 0.982 0.980 0.020
#> SRR1274563 1 0.141 0.982 0.980 0.020
#> SRR1274564 2 0.518 0.879 0.116 0.884
#> SRR1274565 2 0.000 0.977 0.000 1.000
#> SRR1274566 2 0.443 0.904 0.092 0.908
#> SRR1274567 2 0.443 0.904 0.092 0.908
#> SRR1274568 2 0.000 0.977 0.000 1.000
#> SRR1274569 1 0.163 0.978 0.976 0.024
#> SRR1274570 1 0.141 0.982 0.980 0.020
#> SRR1274571 1 0.141 0.982 0.980 0.020
#> SRR1274572 1 0.141 0.982 0.980 0.020
#> SRR1274573 2 0.000 0.977 0.000 1.000
#> SRR1274574 2 0.000 0.977 0.000 1.000
#> SRR1274575 2 0.000 0.977 0.000 1.000
#> SRR1274576 2 0.000 0.977 0.000 1.000
#> SRR1274577 2 0.000 0.977 0.000 1.000
#> SRR1274578 1 0.141 0.982 0.980 0.020
#> SRR1274579 1 0.141 0.982 0.980 0.020
#> SRR1274580 1 0.141 0.982 0.980 0.020
#> SRR1274581 1 0.000 0.963 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274531 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274532 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274533 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274534 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274535 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274536 2 0.3425 0.869 0.112 0.884 0.004
#> SRR1274537 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274539 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274549 2 0.5201 0.691 0.236 0.760 0.004
#> SRR1274550 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274551 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274552 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274553 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274554 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274555 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274556 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274557 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274558 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274559 2 0.2356 0.914 0.072 0.928 0.000
#> SRR1274560 2 0.2356 0.914 0.072 0.928 0.000
#> SRR1274561 1 0.5138 0.499 0.748 0.252 0.000
#> SRR1274562 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274563 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274564 2 0.3425 0.869 0.112 0.884 0.004
#> SRR1274565 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274566 2 0.2945 0.895 0.088 0.908 0.004
#> SRR1274567 2 0.2945 0.895 0.088 0.908 0.004
#> SRR1274568 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274569 1 0.0237 0.964 0.996 0.004 0.000
#> SRR1274570 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274577 2 0.0000 0.974 0.000 1.000 0.000
#> SRR1274578 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.970 1.000 0.000 0.000
#> SRR1274581 3 0.0000 0.000 0.000 0.000 1.000
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274529 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274530 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274531 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274532 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274533 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274534 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274535 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274536 2 0.4282 0.857 0.124 0.816 0.060 0
#> SRR1274537 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274538 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274539 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274540 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274541 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274542 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274543 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274544 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274545 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274546 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274547 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274548 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274549 2 0.5952 0.694 0.124 0.692 0.184 0
#> SRR1274550 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274551 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274552 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274553 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274554 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274555 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274556 3 0.0188 0.947 0.004 0.000 0.996 0
#> SRR1274557 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274558 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274559 2 0.3280 0.891 0.124 0.860 0.016 0
#> SRR1274560 2 0.3280 0.891 0.124 0.860 0.016 0
#> SRR1274561 1 0.7372 0.347 0.524 0.236 0.240 0
#> SRR1274562 1 0.4382 0.693 0.704 0.000 0.296 0
#> SRR1274563 3 0.1022 0.991 0.032 0.000 0.968 0
#> SRR1274564 2 0.4282 0.857 0.124 0.816 0.060 0
#> SRR1274565 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274566 2 0.3821 0.879 0.120 0.840 0.040 0
#> SRR1274567 2 0.3821 0.879 0.120 0.840 0.040 0
#> SRR1274568 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274569 1 0.2589 0.889 0.884 0.000 0.116 0
#> SRR1274570 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274571 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274572 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274573 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274574 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274575 2 0.0000 0.950 0.000 1.000 0.000 0
#> SRR1274576 2 0.0336 0.949 0.008 0.992 0.000 0
#> SRR1274577 2 0.1792 0.938 0.068 0.932 0.000 0
#> SRR1274578 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274579 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274580 1 0.2647 0.893 0.880 0.000 0.120 0
#> SRR1274581 4 0.0000 0.000 0.000 0.000 0.000 1
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274529 2 0.0162 0.896 0.000 0.996 0.000 0.004 0
#> SRR1274530 2 0.0162 0.896 0.000 0.996 0.000 0.004 0
#> SRR1274531 2 0.2674 0.806 0.004 0.856 0.000 0.140 0
#> SRR1274532 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274533 2 0.2179 0.851 0.004 0.896 0.000 0.100 0
#> SRR1274534 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274535 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274536 4 0.4307 0.381 0.000 0.500 0.000 0.500 0
#> SRR1274537 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274538 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274539 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274540 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274541 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274542 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274543 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274544 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274545 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274546 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274547 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274548 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274549 4 0.5236 0.490 0.000 0.380 0.052 0.568 0
#> SRR1274550 3 0.0000 0.997 0.000 0.000 1.000 0.000 0
#> SRR1274551 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274552 3 0.0000 0.997 0.000 0.000 1.000 0.000 0
#> SRR1274553 3 0.0000 0.997 0.000 0.000 1.000 0.000 0
#> SRR1274554 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274555 3 0.0162 0.997 0.000 0.000 0.996 0.004 0
#> SRR1274556 4 0.4287 -0.603 0.000 0.000 0.460 0.540 0
#> SRR1274557 3 0.0162 0.997 0.000 0.000 0.996 0.004 0
#> SRR1274558 2 0.1892 0.862 0.004 0.916 0.000 0.080 0
#> SRR1274559 2 0.3840 0.732 0.076 0.808 0.000 0.116 0
#> SRR1274560 2 0.3840 0.732 0.076 0.808 0.000 0.116 0
#> SRR1274561 1 0.7075 0.280 0.536 0.200 0.212 0.052 0
#> SRR1274562 1 0.3934 0.568 0.716 0.000 0.276 0.008 0
#> SRR1274563 3 0.0162 0.997 0.000 0.000 0.996 0.004 0
#> SRR1274564 4 0.4307 0.381 0.000 0.500 0.000 0.500 0
#> SRR1274565 2 0.2674 0.806 0.004 0.856 0.000 0.140 0
#> SRR1274566 2 0.4074 0.078 0.000 0.636 0.000 0.364 0
#> SRR1274567 2 0.4074 0.078 0.000 0.636 0.000 0.364 0
#> SRR1274568 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274569 1 0.0000 0.871 1.000 0.000 0.000 0.000 0
#> SRR1274570 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274571 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274572 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274573 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274574 2 0.0000 0.899 0.000 1.000 0.000 0.000 0
#> SRR1274575 2 0.0162 0.896 0.000 0.996 0.000 0.004 0
#> SRR1274576 2 0.0955 0.887 0.004 0.968 0.000 0.028 0
#> SRR1274577 2 0.2124 0.853 0.004 0.900 0.000 0.096 0
#> SRR1274578 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274579 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274580 1 0.0162 0.875 0.996 0.000 0.004 0.000 0
#> SRR1274581 5 0.0000 0.000 0.000 0.000 0.000 0.000 1
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274529 2 0.1204 0.866 0.000 0.944 0.000 0.056 0.000 0
#> SRR1274530 2 0.1204 0.866 0.000 0.944 0.000 0.056 0.000 0
#> SRR1274531 2 0.4046 0.793 0.000 0.748 0.000 0.168 0.084 0
#> SRR1274532 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274533 2 0.3509 0.842 0.000 0.804 0.000 0.112 0.084 0
#> SRR1274534 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274535 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274536 4 0.1531 0.769 0.000 0.068 0.000 0.928 0.004 0
#> SRR1274537 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274538 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274539 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274540 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274541 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274542 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274543 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274544 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274545 2 0.0937 0.877 0.000 0.960 0.000 0.040 0.000 0
#> SRR1274546 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274547 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274548 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274549 4 0.2510 0.600 0.000 0.000 0.028 0.872 0.100 0
#> SRR1274550 3 0.0000 0.992 0.000 0.000 1.000 0.000 0.000 0
#> SRR1274551 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274552 3 0.0000 0.992 0.000 0.000 1.000 0.000 0.000 0
#> SRR1274553 3 0.0000 0.992 0.000 0.000 1.000 0.000 0.000 0
#> SRR1274554 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274555 3 0.0363 0.992 0.000 0.000 0.988 0.000 0.012 0
#> SRR1274556 5 0.1802 0.000 0.000 0.000 0.072 0.012 0.916 0
#> SRR1274557 3 0.0363 0.992 0.000 0.000 0.988 0.000 0.012 0
#> SRR1274558 2 0.3125 0.857 0.000 0.836 0.000 0.080 0.084 0
#> SRR1274559 2 0.4809 0.781 0.068 0.736 0.000 0.112 0.084 0
#> SRR1274560 2 0.4809 0.781 0.068 0.736 0.000 0.112 0.084 0
#> SRR1274561 1 0.6654 0.246 0.528 0.196 0.212 0.052 0.012 0
#> SRR1274562 1 0.3652 0.553 0.720 0.000 0.264 0.000 0.016 0
#> SRR1274563 3 0.0363 0.992 0.000 0.000 0.988 0.000 0.012 0
#> SRR1274564 4 0.1531 0.769 0.000 0.068 0.000 0.928 0.004 0
#> SRR1274565 2 0.4046 0.793 0.000 0.748 0.000 0.168 0.084 0
#> SRR1274566 4 0.2793 0.719 0.000 0.200 0.000 0.800 0.000 0
#> SRR1274567 4 0.2793 0.719 0.000 0.200 0.000 0.800 0.000 0
#> SRR1274568 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274569 1 0.0146 0.872 0.996 0.000 0.000 0.000 0.004 0
#> SRR1274570 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274571 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274572 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274573 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274574 2 0.0000 0.901 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274575 2 0.1204 0.866 0.000 0.944 0.000 0.056 0.000 0
#> SRR1274576 2 0.2511 0.872 0.000 0.880 0.000 0.064 0.056 0
#> SRR1274577 2 0.3419 0.846 0.000 0.812 0.000 0.104 0.084 0
#> SRR1274578 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274579 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274580 1 0.0000 0.875 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274581 6 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 1
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "kmeans"]
# you can also extract it by
# res = res_list["MAD:kmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'kmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.983 0.990 0.4783 0.525 0.525
#> 3 3 0.517 0.249 0.660 0.3028 0.900 0.817
#> 4 4 0.612 0.657 0.791 0.1507 0.708 0.434
#> 5 5 0.767 0.748 0.837 0.0761 0.867 0.563
#> 6 6 0.809 0.687 0.806 0.0443 0.957 0.801
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.983 0.000 1.000
#> SRR1274529 2 0.0000 0.983 0.000 1.000
#> SRR1274530 2 0.0000 0.983 0.000 1.000
#> SRR1274531 2 0.3733 0.938 0.072 0.928
#> SRR1274532 2 0.0000 0.983 0.000 1.000
#> SRR1274533 2 0.1843 0.967 0.028 0.972
#> SRR1274534 2 0.0376 0.981 0.004 0.996
#> SRR1274535 2 0.3733 0.938 0.072 0.928
#> SRR1274536 2 0.3733 0.938 0.072 0.928
#> SRR1274537 2 0.0000 0.983 0.000 1.000
#> SRR1274538 2 0.0000 0.983 0.000 1.000
#> SRR1274539 2 0.0376 0.981 0.004 0.996
#> SRR1274540 2 0.0000 0.983 0.000 1.000
#> SRR1274541 2 0.0000 0.983 0.000 1.000
#> SRR1274542 2 0.0000 0.983 0.000 1.000
#> SRR1274543 2 0.0000 0.983 0.000 1.000
#> SRR1274544 2 0.0000 0.983 0.000 1.000
#> SRR1274545 2 0.0000 0.983 0.000 1.000
#> SRR1274546 2 0.0000 0.983 0.000 1.000
#> SRR1274547 2 0.0000 0.983 0.000 1.000
#> SRR1274548 2 0.0000 0.983 0.000 1.000
#> SRR1274549 1 0.0000 1.000 1.000 0.000
#> SRR1274550 1 0.0000 1.000 1.000 0.000
#> SRR1274551 2 0.0000 0.983 0.000 1.000
#> SRR1274552 1 0.0000 1.000 1.000 0.000
#> SRR1274553 1 0.0000 1.000 1.000 0.000
#> SRR1274554 2 0.3733 0.938 0.072 0.928
#> SRR1274555 1 0.0000 1.000 1.000 0.000
#> SRR1274556 1 0.0000 1.000 1.000 0.000
#> SRR1274557 1 0.0000 1.000 1.000 0.000
#> SRR1274558 2 0.0000 0.983 0.000 1.000
#> SRR1274559 1 0.0000 1.000 1.000 0.000
#> SRR1274560 1 0.0000 1.000 1.000 0.000
#> SRR1274561 1 0.0000 1.000 1.000 0.000
#> SRR1274562 1 0.0000 1.000 1.000 0.000
#> SRR1274563 1 0.0000 1.000 1.000 0.000
#> SRR1274564 2 0.0000 0.983 0.000 1.000
#> SRR1274565 2 0.0000 0.983 0.000 1.000
#> SRR1274566 2 0.0000 0.983 0.000 1.000
#> SRR1274567 2 0.4562 0.914 0.096 0.904
#> SRR1274568 2 0.3733 0.938 0.072 0.928
#> SRR1274569 1 0.0000 1.000 1.000 0.000
#> SRR1274570 1 0.0000 1.000 1.000 0.000
#> SRR1274571 1 0.0000 1.000 1.000 0.000
#> SRR1274572 1 0.0000 1.000 1.000 0.000
#> SRR1274573 2 0.0000 0.983 0.000 1.000
#> SRR1274574 2 0.0000 0.983 0.000 1.000
#> SRR1274575 2 0.0000 0.983 0.000 1.000
#> SRR1274576 2 0.0000 0.983 0.000 1.000
#> SRR1274577 2 0.3733 0.938 0.072 0.928
#> SRR1274578 1 0.0000 1.000 1.000 0.000
#> SRR1274579 1 0.0000 1.000 1.000 0.000
#> SRR1274580 1 0.0000 1.000 1.000 0.000
#> SRR1274581 1 0.0000 1.000 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274529 3 0.628 1.0000 0.000 0.460 0.540
#> SRR1274530 3 0.628 1.0000 0.000 0.460 0.540
#> SRR1274531 2 0.826 0.2115 0.180 0.636 0.184
#> SRR1274532 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274533 2 0.472 0.2901 0.160 0.824 0.016
#> SRR1274534 2 0.457 0.2892 0.160 0.828 0.012
#> SRR1274535 2 0.517 0.2805 0.192 0.792 0.016
#> SRR1274536 2 0.807 0.1979 0.160 0.652 0.188
#> SRR1274537 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274538 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274539 2 0.457 0.2892 0.160 0.828 0.012
#> SRR1274540 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274541 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274542 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274543 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274544 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274545 2 0.621 -0.6009 0.000 0.572 0.428
#> SRR1274546 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274547 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274548 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274549 1 0.853 0.4401 0.604 0.240 0.156
#> SRR1274550 1 0.186 0.7533 0.948 0.052 0.000
#> SRR1274551 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274552 1 0.364 0.7140 0.872 0.124 0.004
#> SRR1274553 1 0.210 0.7539 0.944 0.052 0.004
#> SRR1274554 2 0.472 0.2901 0.160 0.824 0.016
#> SRR1274555 1 0.000 0.7672 1.000 0.000 0.000
#> SRR1274556 1 0.165 0.7598 0.960 0.004 0.036
#> SRR1274557 1 0.341 0.7136 0.876 0.124 0.000
#> SRR1274558 2 0.304 -0.0522 0.000 0.896 0.104
#> SRR1274559 2 0.965 -0.4098 0.380 0.412 0.208
#> SRR1274560 1 0.990 0.4610 0.380 0.352 0.268
#> SRR1274561 1 0.645 0.7128 0.760 0.152 0.088
#> SRR1274562 1 0.475 0.7738 0.784 0.000 0.216
#> SRR1274563 1 0.000 0.7672 1.000 0.000 0.000
#> SRR1274564 2 0.435 0.0598 0.000 0.816 0.184
#> SRR1274565 2 0.429 0.0588 0.000 0.820 0.180
#> SRR1274566 2 0.435 0.0598 0.000 0.816 0.184
#> SRR1274567 2 0.812 0.1971 0.164 0.648 0.188
#> SRR1274568 2 0.434 0.2795 0.136 0.848 0.016
#> SRR1274569 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274570 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274571 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274572 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274573 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274574 2 0.595 -0.3152 0.000 0.640 0.360
#> SRR1274575 3 0.628 1.0000 0.000 0.460 0.540
#> SRR1274576 2 0.599 -0.3188 0.000 0.632 0.368
#> SRR1274577 2 0.472 0.2901 0.160 0.824 0.016
#> SRR1274578 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274579 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274580 1 0.617 0.7651 0.588 0.000 0.412
#> SRR1274581 1 0.493 0.7641 0.768 0.000 0.232
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.5268 0.45149 0.000 0.592 0.012 0.396
#> SRR1274530 2 0.5268 0.45149 0.000 0.592 0.012 0.396
#> SRR1274531 4 0.6524 0.61049 0.004 0.092 0.296 0.608
#> SRR1274532 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.7869 0.64776 0.004 0.228 0.348 0.420
#> SRR1274534 4 0.7723 0.64645 0.000 0.232 0.348 0.420
#> SRR1274535 4 0.7763 0.62244 0.004 0.200 0.376 0.420
#> SRR1274536 4 0.2998 0.54596 0.004 0.080 0.024 0.892
#> SRR1274537 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274539 4 0.7723 0.64645 0.000 0.232 0.348 0.420
#> SRR1274540 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.3377 0.75093 0.000 0.848 0.012 0.140
#> SRR1274546 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0804 0.87201 0.000 0.980 0.012 0.008
#> SRR1274548 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.5980 0.43759 0.044 0.000 0.560 0.396
#> SRR1274550 3 0.3217 0.79571 0.128 0.000 0.860 0.012
#> SRR1274551 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.2699 0.73033 0.068 0.000 0.904 0.028
#> SRR1274553 3 0.3653 0.78819 0.128 0.000 0.844 0.028
#> SRR1274554 4 0.7869 0.64776 0.004 0.228 0.348 0.420
#> SRR1274555 3 0.4391 0.71171 0.252 0.000 0.740 0.008
#> SRR1274556 3 0.4327 0.71616 0.216 0.000 0.768 0.016
#> SRR1274557 3 0.3217 0.79665 0.128 0.000 0.860 0.012
#> SRR1274558 2 0.6694 -0.14122 0.004 0.560 0.088 0.348
#> SRR1274559 4 0.8029 0.35840 0.332 0.016 0.200 0.452
#> SRR1274560 4 0.7755 0.29538 0.368 0.004 0.200 0.428
#> SRR1274561 1 0.7723 -0.13536 0.420 0.000 0.348 0.232
#> SRR1274562 1 0.5294 -0.10969 0.508 0.000 0.484 0.008
#> SRR1274563 3 0.4391 0.71171 0.252 0.000 0.740 0.008
#> SRR1274564 4 0.2859 0.54402 0.000 0.112 0.008 0.880
#> SRR1274565 4 0.2888 0.55049 0.000 0.124 0.004 0.872
#> SRR1274566 4 0.2983 0.54535 0.004 0.108 0.008 0.880
#> SRR1274567 4 0.2998 0.54596 0.004 0.080 0.024 0.892
#> SRR1274568 4 0.7815 0.64386 0.004 0.232 0.312 0.452
#> SRR1274569 1 0.0921 0.77701 0.972 0.000 0.000 0.028
#> SRR1274570 1 0.0921 0.77701 0.972 0.000 0.000 0.028
#> SRR1274571 1 0.0921 0.77701 0.972 0.000 0.000 0.028
#> SRR1274572 1 0.0000 0.77909 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.88549 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.5268 0.45149 0.000 0.592 0.012 0.396
#> SRR1274576 2 0.0779 0.86697 0.000 0.980 0.004 0.016
#> SRR1274577 4 0.7869 0.64776 0.004 0.228 0.348 0.420
#> SRR1274578 1 0.0336 0.77922 0.992 0.000 0.008 0.000
#> SRR1274579 1 0.0336 0.77922 0.992 0.000 0.008 0.000
#> SRR1274580 1 0.0336 0.77922 0.992 0.000 0.008 0.000
#> SRR1274581 1 0.6134 0.00849 0.508 0.000 0.444 0.048
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 4 0.5900 0.359 0.000 0.448 0.076 0.468 0.008
#> SRR1274530 4 0.5900 0.359 0.000 0.448 0.076 0.468 0.008
#> SRR1274531 5 0.4026 0.587 0.000 0.020 0.000 0.244 0.736
#> SRR1274532 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.1956 0.798 0.000 0.076 0.008 0.000 0.916
#> SRR1274534 5 0.1956 0.798 0.000 0.076 0.008 0.000 0.916
#> SRR1274535 5 0.1800 0.778 0.000 0.048 0.020 0.000 0.932
#> SRR1274536 4 0.3841 0.646 0.000 0.032 0.012 0.808 0.148
#> SRR1274537 2 0.0162 0.965 0.000 0.996 0.004 0.000 0.000
#> SRR1274538 2 0.0162 0.965 0.000 0.996 0.004 0.000 0.000
#> SRR1274539 5 0.1956 0.798 0.000 0.076 0.008 0.000 0.916
#> SRR1274540 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0162 0.965 0.000 0.996 0.004 0.000 0.000
#> SRR1274544 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.3466 0.755 0.000 0.844 0.048 0.100 0.008
#> SRR1274546 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0798 0.944 0.000 0.976 0.016 0.000 0.008
#> SRR1274548 2 0.0162 0.965 0.000 0.996 0.004 0.000 0.000
#> SRR1274549 3 0.5613 0.352 0.004 0.000 0.520 0.412 0.064
#> SRR1274550 3 0.3913 0.735 0.032 0.000 0.824 0.036 0.108
#> SRR1274551 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.5368 0.590 0.016 0.000 0.632 0.048 0.304
#> SRR1274553 3 0.5619 0.612 0.032 0.000 0.632 0.048 0.288
#> SRR1274554 5 0.1956 0.798 0.000 0.076 0.008 0.000 0.916
#> SRR1274555 3 0.2351 0.734 0.088 0.000 0.896 0.000 0.016
#> SRR1274556 3 0.2541 0.734 0.068 0.000 0.900 0.020 0.012
#> SRR1274557 3 0.2740 0.740 0.028 0.000 0.876 0.000 0.096
#> SRR1274558 5 0.5215 0.294 0.000 0.444 0.008 0.028 0.520
#> SRR1274559 5 0.6688 0.410 0.244 0.000 0.016 0.212 0.528
#> SRR1274560 5 0.6660 0.392 0.264 0.000 0.016 0.192 0.528
#> SRR1274561 3 0.7651 0.365 0.236 0.000 0.444 0.252 0.068
#> SRR1274562 3 0.4322 0.583 0.260 0.000 0.716 0.012 0.012
#> SRR1274563 3 0.2351 0.734 0.088 0.000 0.896 0.000 0.016
#> SRR1274564 4 0.3608 0.653 0.000 0.040 0.000 0.812 0.148
#> SRR1274565 4 0.3994 0.618 0.000 0.040 0.000 0.772 0.188
#> SRR1274566 4 0.3608 0.653 0.000 0.040 0.000 0.812 0.148
#> SRR1274567 4 0.3841 0.646 0.000 0.032 0.012 0.808 0.148
#> SRR1274568 5 0.3058 0.779 0.008 0.076 0.004 0.036 0.876
#> SRR1274569 1 0.2321 0.920 0.912 0.000 0.008 0.056 0.024
#> SRR1274570 1 0.2251 0.921 0.916 0.000 0.008 0.052 0.024
#> SRR1274571 1 0.2321 0.920 0.912 0.000 0.008 0.056 0.024
#> SRR1274572 1 0.0404 0.938 0.988 0.000 0.012 0.000 0.000
#> SRR1274573 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.966 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.5900 0.359 0.000 0.448 0.076 0.468 0.008
#> SRR1274576 2 0.3398 0.660 0.000 0.780 0.004 0.000 0.216
#> SRR1274577 5 0.1956 0.798 0.000 0.076 0.008 0.000 0.916
#> SRR1274578 1 0.0968 0.936 0.972 0.000 0.012 0.012 0.004
#> SRR1274579 1 0.0968 0.936 0.972 0.000 0.012 0.012 0.004
#> SRR1274580 1 0.0968 0.936 0.972 0.000 0.012 0.012 0.004
#> SRR1274581 3 0.7030 0.297 0.324 0.000 0.484 0.152 0.040
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.6256 0.4406 0.000 0.292 0.004 0.428 0.004 0.272
#> SRR1274530 4 0.6256 0.4406 0.000 0.292 0.004 0.428 0.004 0.272
#> SRR1274531 5 0.3929 0.3814 0.000 0.000 0.000 0.272 0.700 0.028
#> SRR1274532 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0405 0.8048 0.000 0.008 0.004 0.000 0.988 0.000
#> SRR1274534 5 0.0405 0.8048 0.000 0.008 0.004 0.000 0.988 0.000
#> SRR1274535 5 0.0405 0.8010 0.000 0.004 0.008 0.000 0.988 0.000
#> SRR1274536 4 0.1606 0.5840 0.000 0.004 0.000 0.932 0.056 0.008
#> SRR1274537 2 0.0291 0.9425 0.000 0.992 0.004 0.000 0.000 0.004
#> SRR1274538 2 0.0405 0.9417 0.000 0.988 0.004 0.000 0.000 0.008
#> SRR1274539 5 0.0405 0.8048 0.000 0.008 0.004 0.000 0.988 0.000
#> SRR1274540 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0405 0.9417 0.000 0.988 0.004 0.000 0.000 0.008
#> SRR1274544 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.4098 0.6353 0.000 0.732 0.008 0.032 0.004 0.224
#> SRR1274546 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.1493 0.8960 0.000 0.936 0.004 0.000 0.004 0.056
#> SRR1274548 2 0.0405 0.9417 0.000 0.988 0.004 0.000 0.000 0.008
#> SRR1274549 4 0.4389 -0.0685 0.000 0.000 0.444 0.536 0.012 0.008
#> SRR1274550 3 0.4115 0.7587 0.016 0.000 0.796 0.020 0.064 0.104
#> SRR1274551 2 0.0000 0.9453 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 3 0.6113 0.5824 0.016 0.000 0.564 0.020 0.256 0.144
#> SRR1274553 3 0.6116 0.5997 0.020 0.000 0.576 0.020 0.240 0.144
#> SRR1274554 5 0.0972 0.7984 0.000 0.008 0.000 0.000 0.964 0.028
#> SRR1274555 3 0.0632 0.8037 0.024 0.000 0.976 0.000 0.000 0.000
#> SRR1274556 3 0.1026 0.7974 0.008 0.000 0.968 0.012 0.008 0.004
#> SRR1274557 3 0.1151 0.8032 0.012 0.000 0.956 0.000 0.032 0.000
#> SRR1274558 5 0.5998 0.0948 0.008 0.428 0.004 0.016 0.448 0.096
#> SRR1274559 6 0.7344 0.6894 0.116 0.000 0.008 0.164 0.316 0.396
#> SRR1274560 6 0.7351 0.6898 0.120 0.000 0.008 0.160 0.316 0.396
#> SRR1274561 6 0.7416 0.3699 0.104 0.000 0.288 0.212 0.008 0.388
#> SRR1274562 3 0.3078 0.6789 0.108 0.000 0.836 0.000 0.000 0.056
#> SRR1274563 3 0.0632 0.8037 0.024 0.000 0.976 0.000 0.000 0.000
#> SRR1274564 4 0.1606 0.5872 0.000 0.004 0.000 0.932 0.056 0.008
#> SRR1274565 4 0.2804 0.5164 0.000 0.004 0.000 0.852 0.120 0.024
#> SRR1274566 4 0.1349 0.5872 0.000 0.004 0.000 0.940 0.056 0.000
#> SRR1274567 4 0.1349 0.5872 0.000 0.004 0.000 0.940 0.056 0.000
#> SRR1274568 5 0.2742 0.7232 0.016 0.008 0.000 0.020 0.880 0.076
#> SRR1274569 1 0.3774 0.5220 0.664 0.000 0.008 0.000 0.000 0.328
#> SRR1274570 1 0.3758 0.5259 0.668 0.000 0.008 0.000 0.000 0.324
#> SRR1274571 1 0.3774 0.5220 0.664 0.000 0.008 0.000 0.000 0.328
#> SRR1274572 1 0.1616 0.6909 0.932 0.000 0.020 0.000 0.000 0.048
#> SRR1274573 2 0.0146 0.9440 0.000 0.996 0.004 0.000 0.000 0.000
#> SRR1274574 2 0.0146 0.9440 0.000 0.996 0.004 0.000 0.000 0.000
#> SRR1274575 4 0.6256 0.4406 0.000 0.292 0.004 0.428 0.004 0.272
#> SRR1274576 2 0.4255 0.3044 0.000 0.600 0.004 0.000 0.380 0.016
#> SRR1274577 5 0.0972 0.7984 0.000 0.008 0.000 0.000 0.964 0.028
#> SRR1274578 1 0.0547 0.6969 0.980 0.000 0.020 0.000 0.000 0.000
#> SRR1274579 1 0.0547 0.6969 0.980 0.000 0.020 0.000 0.000 0.000
#> SRR1274580 1 0.0547 0.6969 0.980 0.000 0.020 0.000 0.000 0.000
#> SRR1274581 1 0.6970 -0.0270 0.392 0.000 0.336 0.076 0.000 0.196
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "skmeans"]
# you can also extract it by
# res = res_list["MAD:skmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'skmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.967 0.986 0.5092 0.491 0.491
#> 3 3 0.673 0.659 0.827 0.2770 0.837 0.678
#> 4 4 0.824 0.861 0.939 0.1388 0.820 0.544
#> 5 5 0.882 0.846 0.925 0.0659 0.911 0.676
#> 6 6 0.862 0.783 0.890 0.0315 0.945 0.759
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.999 0.000 1.000
#> SRR1274529 2 0.0000 0.999 0.000 1.000
#> SRR1274530 2 0.0000 0.999 0.000 1.000
#> SRR1274531 1 0.5946 0.830 0.856 0.144
#> SRR1274532 2 0.0000 0.999 0.000 1.000
#> SRR1274533 2 0.1414 0.979 0.020 0.980
#> SRR1274534 2 0.0000 0.999 0.000 1.000
#> SRR1274535 1 0.0000 0.972 1.000 0.000
#> SRR1274536 1 0.4431 0.890 0.908 0.092
#> SRR1274537 2 0.0000 0.999 0.000 1.000
#> SRR1274538 2 0.0000 0.999 0.000 1.000
#> SRR1274539 2 0.0000 0.999 0.000 1.000
#> SRR1274540 2 0.0000 0.999 0.000 1.000
#> SRR1274541 2 0.0000 0.999 0.000 1.000
#> SRR1274542 2 0.0000 0.999 0.000 1.000
#> SRR1274543 2 0.0000 0.999 0.000 1.000
#> SRR1274544 2 0.0000 0.999 0.000 1.000
#> SRR1274545 2 0.0000 0.999 0.000 1.000
#> SRR1274546 2 0.0000 0.999 0.000 1.000
#> SRR1274547 2 0.0000 0.999 0.000 1.000
#> SRR1274548 2 0.0000 0.999 0.000 1.000
#> SRR1274549 1 0.0000 0.972 1.000 0.000
#> SRR1274550 1 0.0000 0.972 1.000 0.000
#> SRR1274551 2 0.0000 0.999 0.000 1.000
#> SRR1274552 1 0.0000 0.972 1.000 0.000
#> SRR1274553 1 0.0000 0.972 1.000 0.000
#> SRR1274554 1 0.9850 0.260 0.572 0.428
#> SRR1274555 1 0.0000 0.972 1.000 0.000
#> SRR1274556 1 0.0000 0.972 1.000 0.000
#> SRR1274557 1 0.0000 0.972 1.000 0.000
#> SRR1274558 2 0.0000 0.999 0.000 1.000
#> SRR1274559 1 0.0000 0.972 1.000 0.000
#> SRR1274560 1 0.0000 0.972 1.000 0.000
#> SRR1274561 1 0.0000 0.972 1.000 0.000
#> SRR1274562 1 0.0000 0.972 1.000 0.000
#> SRR1274563 1 0.0000 0.972 1.000 0.000
#> SRR1274564 2 0.0000 0.999 0.000 1.000
#> SRR1274565 2 0.0000 0.999 0.000 1.000
#> SRR1274566 2 0.0000 0.999 0.000 1.000
#> SRR1274567 1 0.0000 0.972 1.000 0.000
#> SRR1274568 1 0.2948 0.930 0.948 0.052
#> SRR1274569 1 0.0000 0.972 1.000 0.000
#> SRR1274570 1 0.0000 0.972 1.000 0.000
#> SRR1274571 1 0.0000 0.972 1.000 0.000
#> SRR1274572 1 0.0000 0.972 1.000 0.000
#> SRR1274573 2 0.0000 0.999 0.000 1.000
#> SRR1274574 2 0.0000 0.999 0.000 1.000
#> SRR1274575 2 0.0000 0.999 0.000 1.000
#> SRR1274576 2 0.0000 0.999 0.000 1.000
#> SRR1274577 1 0.0376 0.969 0.996 0.004
#> SRR1274578 1 0.0000 0.972 1.000 0.000
#> SRR1274579 1 0.0000 0.972 1.000 0.000
#> SRR1274580 1 0.0000 0.972 1.000 0.000
#> SRR1274581 1 0.0000 0.972 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274529 2 0.5465 0.683 0.000 0.712 0.288
#> SRR1274530 2 0.5465 0.683 0.000 0.712 0.288
#> SRR1274531 3 0.0000 0.528 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274533 3 0.5465 0.597 0.000 0.288 0.712
#> SRR1274534 3 0.5465 0.597 0.000 0.288 0.712
#> SRR1274535 3 0.5465 0.342 0.288 0.000 0.712
#> SRR1274536 3 0.6587 0.250 0.352 0.016 0.632
#> SRR1274537 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274539 3 0.5465 0.597 0.000 0.288 0.712
#> SRR1274540 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274545 2 0.1643 0.872 0.000 0.956 0.044
#> SRR1274546 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274549 3 0.5988 0.222 0.368 0.000 0.632
#> SRR1274550 1 0.6302 0.207 0.520 0.000 0.480
#> SRR1274551 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274552 1 0.6302 0.207 0.520 0.000 0.480
#> SRR1274553 1 0.6302 0.207 0.520 0.000 0.480
#> SRR1274554 3 0.6853 0.589 0.064 0.224 0.712
#> SRR1274555 1 0.5327 0.527 0.728 0.000 0.272
#> SRR1274556 1 0.5327 0.527 0.728 0.000 0.272
#> SRR1274557 1 0.6302 0.207 0.520 0.000 0.480
#> SRR1274558 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274559 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274560 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274561 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274562 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274563 1 0.5327 0.527 0.728 0.000 0.272
#> SRR1274564 2 0.5968 0.603 0.000 0.636 0.364
#> SRR1274565 2 0.6026 0.589 0.000 0.624 0.376
#> SRR1274566 2 0.6026 0.589 0.000 0.624 0.376
#> SRR1274567 3 0.6252 0.090 0.444 0.000 0.556
#> SRR1274568 1 0.9290 -0.142 0.464 0.164 0.372
#> SRR1274569 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.898 0.000 1.000 0.000
#> SRR1274575 2 0.5465 0.683 0.000 0.712 0.288
#> SRR1274576 2 0.0892 0.881 0.000 0.980 0.020
#> SRR1274577 3 0.6703 0.438 0.236 0.052 0.712
#> SRR1274578 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.761 1.000 0.000 0.000
#> SRR1274581 1 0.0000 0.761 1.000 0.000 0.000
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274529 4 0.2469 0.877 0.000 0.108 0.000 0.892
#> SRR1274530 4 0.2469 0.877 0.000 0.108 0.000 0.892
#> SRR1274531 3 0.4500 0.445 0.000 0.000 0.684 0.316
#> SRR1274532 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274534 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274535 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274536 4 0.0000 0.918 0.000 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274539 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0469 0.987 0.000 0.988 0.000 0.012
#> SRR1274546 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274549 4 0.4581 0.705 0.120 0.000 0.080 0.800
#> SRR1274550 3 0.4193 0.655 0.268 0.000 0.732 0.000
#> SRR1274551 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.4222 0.650 0.272 0.000 0.728 0.000
#> SRR1274553 3 0.4356 0.618 0.292 0.000 0.708 0.000
#> SRR1274554 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274555 1 0.4564 0.478 0.672 0.000 0.328 0.000
#> SRR1274556 1 0.4564 0.478 0.672 0.000 0.328 0.000
#> SRR1274557 3 0.4222 0.651 0.272 0.000 0.728 0.000
#> SRR1274558 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274559 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274562 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274563 1 0.4543 0.486 0.676 0.000 0.324 0.000
#> SRR1274564 4 0.0000 0.918 0.000 0.000 0.000 1.000
#> SRR1274565 4 0.0000 0.918 0.000 0.000 0.000 1.000
#> SRR1274566 4 0.0000 0.918 0.000 0.000 0.000 1.000
#> SRR1274567 4 0.0000 0.918 0.000 0.000 0.000 1.000
#> SRR1274568 3 0.5130 0.459 0.312 0.020 0.668 0.000
#> SRR1274569 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.998 0.000 1.000 0.000 0.000
#> SRR1274575 4 0.2469 0.877 0.000 0.108 0.000 0.892
#> SRR1274576 2 0.0817 0.975 0.000 0.976 0.024 0.000
#> SRR1274577 3 0.0000 0.804 0.000 0.000 1.000 0.000
#> SRR1274578 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.909 1.000 0.000 0.000 0.000
#> SRR1274581 1 0.0592 0.896 0.984 0.000 0.016 0.000
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 4 0.2460 0.9160 0.000 0.072 0.024 0.900 0.004
#> SRR1274530 4 0.2460 0.9160 0.000 0.072 0.024 0.900 0.004
#> SRR1274531 3 0.5847 0.0739 0.000 0.000 0.480 0.096 0.424
#> SRR1274532 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.0703 0.9448 0.000 0.000 0.024 0.000 0.976
#> SRR1274534 5 0.0703 0.9448 0.000 0.000 0.024 0.000 0.976
#> SRR1274535 5 0.0880 0.9388 0.000 0.000 0.032 0.000 0.968
#> SRR1274536 4 0.0000 0.9496 0.000 0.000 0.000 1.000 0.000
#> SRR1274537 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.0703 0.9448 0.000 0.000 0.024 0.000 0.976
#> SRR1274540 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.1901 0.9120 0.000 0.932 0.024 0.040 0.004
#> SRR1274546 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0865 0.9463 0.000 0.972 0.024 0.000 0.004
#> SRR1274548 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 3 0.2516 0.7624 0.000 0.000 0.860 0.140 0.000
#> SRR1274550 3 0.1670 0.8363 0.012 0.000 0.936 0.000 0.052
#> SRR1274551 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.1740 0.8347 0.012 0.000 0.932 0.000 0.056
#> SRR1274553 3 0.1670 0.8363 0.012 0.000 0.936 0.000 0.052
#> SRR1274554 5 0.0703 0.9448 0.000 0.000 0.024 0.000 0.976
#> SRR1274555 3 0.0865 0.8338 0.024 0.000 0.972 0.000 0.004
#> SRR1274556 3 0.0865 0.8338 0.024 0.000 0.972 0.000 0.004
#> SRR1274557 3 0.1364 0.8370 0.012 0.000 0.952 0.000 0.036
#> SRR1274558 2 0.2462 0.8574 0.112 0.880 0.000 0.000 0.008
#> SRR1274559 1 0.0290 0.8638 0.992 0.000 0.000 0.000 0.008
#> SRR1274560 1 0.0290 0.8638 0.992 0.000 0.000 0.000 0.008
#> SRR1274561 1 0.4304 -0.0330 0.516 0.000 0.484 0.000 0.000
#> SRR1274562 3 0.4210 0.1455 0.412 0.000 0.588 0.000 0.000
#> SRR1274563 3 0.0865 0.8338 0.024 0.000 0.972 0.000 0.004
#> SRR1274564 4 0.0000 0.9496 0.000 0.000 0.000 1.000 0.000
#> SRR1274565 4 0.0162 0.9481 0.000 0.000 0.000 0.996 0.004
#> SRR1274566 4 0.0000 0.9496 0.000 0.000 0.000 1.000 0.000
#> SRR1274567 4 0.0000 0.9496 0.000 0.000 0.000 1.000 0.000
#> SRR1274568 5 0.3861 0.6335 0.284 0.004 0.000 0.000 0.712
#> SRR1274569 1 0.0703 0.8778 0.976 0.000 0.024 0.000 0.000
#> SRR1274570 1 0.0963 0.8790 0.964 0.000 0.036 0.000 0.000
#> SRR1274571 1 0.0703 0.8778 0.976 0.000 0.024 0.000 0.000
#> SRR1274572 1 0.2179 0.8689 0.888 0.000 0.112 0.000 0.000
#> SRR1274573 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9650 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.2460 0.9160 0.000 0.072 0.024 0.900 0.004
#> SRR1274576 2 0.4088 0.4164 0.000 0.632 0.000 0.000 0.368
#> SRR1274577 5 0.0404 0.9377 0.000 0.000 0.012 0.000 0.988
#> SRR1274578 1 0.2513 0.8662 0.876 0.000 0.116 0.000 0.008
#> SRR1274579 1 0.2513 0.8662 0.876 0.000 0.116 0.000 0.008
#> SRR1274580 1 0.2513 0.8662 0.876 0.000 0.116 0.000 0.008
#> SRR1274581 3 0.3582 0.6121 0.224 0.000 0.768 0.000 0.008
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 4 0.4209 0.8089 0.008 0.080 0.000 0.760 0.004 0.148
#> SRR1274530 4 0.4209 0.8089 0.008 0.080 0.000 0.760 0.004 0.148
#> SRR1274531 3 0.6410 0.0551 0.000 0.000 0.416 0.208 0.352 0.024
#> SRR1274532 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.1196 0.9367 0.000 0.000 0.008 0.000 0.952 0.040
#> SRR1274534 5 0.0146 0.9486 0.000 0.000 0.004 0.000 0.996 0.000
#> SRR1274535 5 0.1649 0.9162 0.000 0.000 0.032 0.000 0.932 0.036
#> SRR1274536 4 0.0000 0.8755 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274537 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.0146 0.9486 0.000 0.000 0.004 0.000 0.996 0.000
#> SRR1274540 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.4052 0.7091 0.008 0.772 0.000 0.068 0.004 0.148
#> SRR1274546 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.2876 0.7902 0.008 0.836 0.000 0.004 0.004 0.148
#> SRR1274548 2 0.0146 0.9457 0.000 0.996 0.000 0.000 0.000 0.004
#> SRR1274549 3 0.2697 0.7166 0.000 0.000 0.812 0.188 0.000 0.000
#> SRR1274550 3 0.1723 0.8164 0.000 0.000 0.928 0.000 0.036 0.036
#> SRR1274551 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 3 0.2265 0.8069 0.000 0.000 0.896 0.000 0.052 0.052
#> SRR1274553 3 0.2265 0.8069 0.000 0.000 0.896 0.000 0.052 0.052
#> SRR1274554 5 0.0935 0.9392 0.000 0.000 0.004 0.000 0.964 0.032
#> SRR1274555 3 0.0000 0.8261 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.0146 0.8258 0.000 0.000 0.996 0.000 0.000 0.004
#> SRR1274557 3 0.0146 0.8261 0.000 0.000 0.996 0.000 0.004 0.000
#> SRR1274558 6 0.4147 0.1590 0.012 0.436 0.000 0.000 0.000 0.552
#> SRR1274559 6 0.3578 0.3991 0.340 0.000 0.000 0.000 0.000 0.660
#> SRR1274560 6 0.3578 0.3991 0.340 0.000 0.000 0.000 0.000 0.660
#> SRR1274561 3 0.4963 0.4469 0.240 0.000 0.636 0.000 0.000 0.124
#> SRR1274562 3 0.3500 0.6373 0.204 0.000 0.768 0.000 0.000 0.028
#> SRR1274563 3 0.0000 0.8261 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 4 0.0790 0.8745 0.000 0.000 0.000 0.968 0.000 0.032
#> SRR1274565 4 0.1075 0.8645 0.000 0.000 0.000 0.952 0.000 0.048
#> SRR1274566 4 0.0458 0.8730 0.000 0.000 0.000 0.984 0.000 0.016
#> SRR1274567 4 0.0458 0.8730 0.000 0.000 0.000 0.984 0.000 0.016
#> SRR1274568 6 0.3885 0.3625 0.044 0.000 0.000 0.000 0.220 0.736
#> SRR1274569 1 0.3043 0.6995 0.792 0.000 0.008 0.000 0.000 0.200
#> SRR1274570 1 0.2948 0.7137 0.804 0.000 0.008 0.000 0.000 0.188
#> SRR1274571 1 0.2948 0.7137 0.804 0.000 0.008 0.000 0.000 0.188
#> SRR1274572 1 0.0891 0.8001 0.968 0.000 0.024 0.000 0.000 0.008
#> SRR1274573 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9487 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 4 0.4209 0.8089 0.008 0.080 0.000 0.760 0.004 0.148
#> SRR1274576 2 0.4640 0.5249 0.000 0.680 0.000 0.000 0.212 0.108
#> SRR1274577 5 0.2191 0.8745 0.004 0.000 0.000 0.000 0.876 0.120
#> SRR1274578 1 0.0632 0.8005 0.976 0.000 0.024 0.000 0.000 0.000
#> SRR1274579 1 0.0632 0.8005 0.976 0.000 0.024 0.000 0.000 0.000
#> SRR1274580 1 0.0632 0.8005 0.976 0.000 0.024 0.000 0.000 0.000
#> SRR1274581 1 0.3861 0.3725 0.640 0.000 0.352 0.000 0.000 0.008
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "pam"]
# you can also extract it by
# res = res_list["MAD:pam"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'pam' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 6.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.887 0.939 0.974 0.5025 0.497 0.497
#> 3 3 0.719 0.697 0.867 0.2247 0.860 0.732
#> 4 4 0.891 0.784 0.895 0.1724 0.809 0.563
#> 5 5 0.848 0.792 0.916 0.0250 0.785 0.425
#> 6 6 0.970 0.896 0.961 0.0716 0.888 0.608
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 6
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.975 0.000 1.000
#> SRR1274529 2 0.000 0.975 0.000 1.000
#> SRR1274530 2 0.000 0.975 0.000 1.000
#> SRR1274531 1 0.552 0.843 0.872 0.128
#> SRR1274532 2 0.000 0.975 0.000 1.000
#> SRR1274533 2 0.000 0.975 0.000 1.000
#> SRR1274534 2 0.000 0.975 0.000 1.000
#> SRR1274535 2 0.866 0.589 0.288 0.712
#> SRR1274536 1 0.552 0.843 0.872 0.128
#> SRR1274537 2 0.000 0.975 0.000 1.000
#> SRR1274538 2 0.000 0.975 0.000 1.000
#> SRR1274539 2 0.000 0.975 0.000 1.000
#> SRR1274540 2 0.000 0.975 0.000 1.000
#> SRR1274541 2 0.000 0.975 0.000 1.000
#> SRR1274542 2 0.000 0.975 0.000 1.000
#> SRR1274543 2 0.000 0.975 0.000 1.000
#> SRR1274544 2 0.000 0.975 0.000 1.000
#> SRR1274545 2 0.000 0.975 0.000 1.000
#> SRR1274546 2 0.000 0.975 0.000 1.000
#> SRR1274547 2 0.000 0.975 0.000 1.000
#> SRR1274548 2 0.000 0.975 0.000 1.000
#> SRR1274549 1 0.000 0.969 1.000 0.000
#> SRR1274550 1 0.000 0.969 1.000 0.000
#> SRR1274551 2 0.000 0.975 0.000 1.000
#> SRR1274552 1 0.000 0.969 1.000 0.000
#> SRR1274553 1 0.000 0.969 1.000 0.000
#> SRR1274554 2 0.373 0.909 0.072 0.928
#> SRR1274555 1 0.000 0.969 1.000 0.000
#> SRR1274556 1 0.000 0.969 1.000 0.000
#> SRR1274557 1 0.000 0.969 1.000 0.000
#> SRR1274558 2 0.000 0.975 0.000 1.000
#> SRR1274559 1 0.000 0.969 1.000 0.000
#> SRR1274560 1 0.000 0.969 1.000 0.000
#> SRR1274561 1 0.000 0.969 1.000 0.000
#> SRR1274562 1 0.000 0.969 1.000 0.000
#> SRR1274563 1 0.000 0.969 1.000 0.000
#> SRR1274564 1 0.990 0.212 0.560 0.440
#> SRR1274565 2 0.000 0.975 0.000 1.000
#> SRR1274566 2 0.714 0.752 0.196 0.804
#> SRR1274567 1 0.000 0.969 1.000 0.000
#> SRR1274568 2 0.563 0.843 0.132 0.868
#> SRR1274569 1 0.000 0.969 1.000 0.000
#> SRR1274570 1 0.000 0.969 1.000 0.000
#> SRR1274571 1 0.000 0.969 1.000 0.000
#> SRR1274572 1 0.000 0.969 1.000 0.000
#> SRR1274573 2 0.000 0.975 0.000 1.000
#> SRR1274574 2 0.000 0.975 0.000 1.000
#> SRR1274575 2 0.000 0.975 0.000 1.000
#> SRR1274576 2 0.000 0.975 0.000 1.000
#> SRR1274577 2 0.000 0.975 0.000 1.000
#> SRR1274578 1 0.000 0.969 1.000 0.000
#> SRR1274579 1 0.000 0.969 1.000 0.000
#> SRR1274580 1 0.000 0.969 1.000 0.000
#> SRR1274581 1 0.000 0.969 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274531 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274533 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274534 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274535 3 0.5650 0.2657 0.000 0.312 0.688
#> SRR1274536 3 0.0237 0.6784 0.000 0.004 0.996
#> SRR1274537 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274539 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274540 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274549 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274550 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274551 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274552 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274553 3 0.1163 0.6725 0.028 0.000 0.972
#> SRR1274554 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274555 3 0.4750 0.5751 0.216 0.000 0.784
#> SRR1274556 3 0.0237 0.6790 0.004 0.000 0.996
#> SRR1274557 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274558 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274559 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274560 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274561 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274562 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274563 3 0.6079 0.4277 0.388 0.000 0.612
#> SRR1274564 3 0.6295 0.2129 0.000 0.472 0.528
#> SRR1274565 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274566 3 0.6154 -0.0528 0.000 0.408 0.592
#> SRR1274567 3 0.0000 0.6793 0.000 0.000 1.000
#> SRR1274568 2 0.6154 0.5037 0.000 0.592 0.408
#> SRR1274569 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274570 1 0.4002 0.7350 0.840 0.000 0.160
#> SRR1274571 3 0.6126 0.4158 0.400 0.000 0.600
#> SRR1274572 1 0.0000 0.9536 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.8812 0.000 1.000 0.000
#> SRR1274577 2 0.6126 0.5183 0.000 0.600 0.400
#> SRR1274578 1 0.0000 0.9536 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.9536 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.9536 1.000 0.000 0.000
#> SRR1274581 1 0.0424 0.9465 0.992 0.000 0.008
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.0469 0.9870 0.000 0.988 0.000 0.012
#> SRR1274531 4 0.0000 0.8781 0.000 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.2081 0.8738 0.000 0.084 0.000 0.916
#> SRR1274534 4 0.2081 0.8738 0.000 0.084 0.000 0.916
#> SRR1274535 4 0.2271 0.8756 0.000 0.076 0.008 0.916
#> SRR1274536 3 0.4855 0.4534 0.000 0.000 0.600 0.400
#> SRR1274537 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274539 4 0.2081 0.8738 0.000 0.084 0.000 0.916
#> SRR1274540 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274549 3 0.2081 0.5722 0.000 0.000 0.916 0.084
#> SRR1274550 3 0.4746 0.0665 0.000 0.000 0.632 0.368
#> SRR1274551 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274552 4 0.2413 0.8505 0.064 0.000 0.020 0.916
#> SRR1274553 4 0.4343 0.6711 0.004 0.000 0.264 0.732
#> SRR1274554 4 0.0000 0.8781 0.000 0.000 0.000 1.000
#> SRR1274555 3 0.0000 0.5653 0.000 0.000 1.000 0.000
#> SRR1274556 3 0.4222 0.3268 0.000 0.000 0.728 0.272
#> SRR1274557 4 0.4855 0.4669 0.000 0.000 0.400 0.600
#> SRR1274558 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274559 3 0.4855 0.4500 0.400 0.000 0.600 0.000
#> SRR1274560 3 0.4855 0.4500 0.400 0.000 0.600 0.000
#> SRR1274561 3 0.4855 0.4500 0.400 0.000 0.600 0.000
#> SRR1274562 3 0.0336 0.5655 0.008 0.000 0.992 0.000
#> SRR1274563 3 0.0000 0.5653 0.000 0.000 1.000 0.000
#> SRR1274564 3 0.6439 0.3401 0.000 0.340 0.576 0.084
#> SRR1274565 4 0.0000 0.8781 0.000 0.000 0.000 1.000
#> SRR1274566 4 0.0000 0.8781 0.000 0.000 0.000 1.000
#> SRR1274567 3 0.4855 0.4534 0.000 0.000 0.600 0.400
#> SRR1274568 4 0.2081 0.8738 0.000 0.084 0.000 0.916
#> SRR1274569 3 0.4855 0.4500 0.400 0.000 0.600 0.000
#> SRR1274570 1 0.3172 0.6497 0.840 0.000 0.160 0.000
#> SRR1274571 3 0.4855 0.4500 0.400 0.000 0.600 0.000
#> SRR1274572 1 0.0000 0.8525 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0469 0.9870 0.000 0.988 0.000 0.012
#> SRR1274576 2 0.0000 0.9986 0.000 1.000 0.000 0.000
#> SRR1274577 4 0.0000 0.8781 0.000 0.000 0.000 1.000
#> SRR1274578 1 0.0000 0.8525 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.8525 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.8525 1.000 0.000 0.000 0.000
#> SRR1274581 1 0.4872 0.4201 0.640 0.000 0.356 0.004
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.0963 0.883 0.000 0.964 0.000 0.036 0.000
#> SRR1274530 4 0.4262 0.355 0.000 0.440 0.000 0.560 0.000
#> SRR1274531 4 0.0000 0.740 0.000 0.000 0.000 1.000 0.000
#> SRR1274532 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 2 0.4384 0.502 0.000 0.660 0.016 0.324 0.000
#> SRR1274534 2 0.4384 0.502 0.000 0.660 0.016 0.324 0.000
#> SRR1274535 3 0.5338 0.489 0.000 0.072 0.604 0.324 0.000
#> SRR1274536 4 0.0162 0.741 0.000 0.000 0.000 0.996 0.004
#> SRR1274537 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 2 0.4384 0.502 0.000 0.660 0.016 0.324 0.000
#> SRR1274540 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 4 0.4384 0.363 0.016 0.000 0.324 0.660 0.000
#> SRR1274550 3 0.0000 0.817 0.000 0.000 1.000 0.000 0.000
#> SRR1274551 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.5172 0.536 0.060 0.000 0.616 0.324 0.000
#> SRR1274553 3 0.3001 0.750 0.008 0.000 0.844 0.144 0.004
#> SRR1274554 4 0.2208 0.679 0.000 0.072 0.020 0.908 0.000
#> SRR1274555 3 0.0510 0.821 0.016 0.000 0.984 0.000 0.000
#> SRR1274556 3 0.0510 0.821 0.016 0.000 0.984 0.000 0.000
#> SRR1274557 3 0.0510 0.821 0.016 0.000 0.984 0.000 0.000
#> SRR1274558 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274559 1 0.0000 0.933 1.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.933 1.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.933 1.000 0.000 0.000 0.000 0.000
#> SRR1274562 1 0.4171 0.391 0.604 0.000 0.396 0.000 0.000
#> SRR1274563 3 0.0510 0.821 0.016 0.000 0.984 0.000 0.000
#> SRR1274564 4 0.4047 0.522 0.000 0.320 0.000 0.676 0.004
#> SRR1274565 4 0.0162 0.741 0.000 0.000 0.000 0.996 0.004
#> SRR1274566 4 0.0162 0.741 0.000 0.000 0.000 0.996 0.004
#> SRR1274567 4 0.0162 0.741 0.000 0.000 0.000 0.996 0.004
#> SRR1274568 2 0.4074 0.453 0.000 0.636 0.000 0.364 0.000
#> SRR1274569 1 0.0000 0.933 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0162 0.931 0.996 0.000 0.000 0.000 0.004
#> SRR1274571 1 0.0000 0.933 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0510 0.919 0.984 0.000 0.000 0.000 0.016
#> SRR1274573 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.4300 0.263 0.000 0.476 0.000 0.524 0.000
#> SRR1274576 2 0.0000 0.919 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 4 0.0510 0.730 0.000 0.000 0.016 0.984 0.000
#> SRR1274578 5 0.0162 0.998 0.004 0.000 0.000 0.000 0.996
#> SRR1274579 5 0.0162 0.998 0.004 0.000 0.000 0.000 0.996
#> SRR1274580 5 0.0162 0.998 0.004 0.000 0.000 0.000 0.996
#> SRR1274581 5 0.0000 0.995 0.000 0.000 0.000 0.000 1.000
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0937 0.952 0.000 0.960 0.000 0.040 0.000 0.000
#> SRR1274530 4 0.3823 0.336 0.000 0.436 0.000 0.564 0.000 0.000
#> SRR1274531 5 0.3838 0.193 0.000 0.000 0.000 0.448 0.552 0.000
#> SRR1274532 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274534 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274535 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274536 4 0.0000 0.767 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274537 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274549 4 0.2454 0.642 0.000 0.000 0.160 0.840 0.000 0.000
#> SRR1274550 3 0.0000 0.929 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274551 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274553 5 0.2946 0.722 0.004 0.000 0.184 0.000 0.808 0.004
#> SRR1274554 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274555 3 0.0000 0.929 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.0000 0.929 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274557 3 0.0000 0.929 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274558 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274559 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274562 3 0.3428 0.565 0.304 0.000 0.696 0.000 0.000 0.000
#> SRR1274563 3 0.0000 0.929 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 4 0.0000 0.767 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274565 4 0.0146 0.768 0.000 0.000 0.000 0.996 0.004 0.000
#> SRR1274566 4 0.0146 0.768 0.000 0.000 0.000 0.996 0.004 0.000
#> SRR1274567 4 0.0146 0.768 0.000 0.000 0.000 0.996 0.004 0.000
#> SRR1274568 5 0.1053 0.893 0.020 0.012 0.000 0.004 0.964 0.000
#> SRR1274569 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 1.000 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 4 0.3860 0.239 0.000 0.472 0.000 0.528 0.000 0.000
#> SRR1274576 2 0.0000 0.997 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 5 0.0000 0.915 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274578 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274579 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274580 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
#> SRR1274581 6 0.0000 1.000 0.000 0.000 0.000 0.000 0.000 1.000
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "mclust"]
# you can also extract it by
# res = res_list["MAD:mclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'mclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 3.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.520 0.814 0.909 0.3564 0.693 0.693
#> 3 3 0.504 0.743 0.862 0.4658 0.756 0.655
#> 4 4 0.471 0.597 0.805 0.2518 0.903 0.805
#> 5 5 0.674 0.674 0.818 0.2036 0.724 0.382
#> 6 6 0.701 0.788 0.831 0.0373 0.913 0.607
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 3
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.9286 0.377 0.344 0.656
#> SRR1274529 1 0.8763 0.657 0.704 0.296
#> SRR1274530 1 0.8763 0.657 0.704 0.296
#> SRR1274531 1 0.1414 0.888 0.980 0.020
#> SRR1274532 2 0.0000 0.911 0.000 1.000
#> SRR1274533 1 0.3114 0.873 0.944 0.056
#> SRR1274534 1 0.0672 0.890 0.992 0.008
#> SRR1274535 1 0.1633 0.887 0.976 0.024
#> SRR1274536 1 0.5842 0.824 0.860 0.140
#> SRR1274537 1 0.9754 0.438 0.592 0.408
#> SRR1274538 1 0.9963 0.279 0.536 0.464
#> SRR1274539 1 0.0672 0.890 0.992 0.008
#> SRR1274540 2 0.0000 0.911 0.000 1.000
#> SRR1274541 2 0.0000 0.911 0.000 1.000
#> SRR1274542 2 0.0000 0.911 0.000 1.000
#> SRR1274543 1 0.8144 0.716 0.748 0.252
#> SRR1274544 2 0.0000 0.911 0.000 1.000
#> SRR1274545 1 0.8909 0.638 0.692 0.308
#> SRR1274546 2 0.4690 0.837 0.100 0.900
#> SRR1274547 1 0.9732 0.448 0.596 0.404
#> SRR1274548 1 0.9393 0.546 0.644 0.356
#> SRR1274549 1 0.1414 0.888 0.980 0.020
#> SRR1274550 1 0.0376 0.891 0.996 0.004
#> SRR1274551 2 0.0000 0.911 0.000 1.000
#> SRR1274552 1 0.0376 0.891 0.996 0.004
#> SRR1274553 1 0.0000 0.890 1.000 0.000
#> SRR1274554 1 0.3114 0.873 0.944 0.056
#> SRR1274555 1 0.0376 0.891 0.996 0.004
#> SRR1274556 1 0.1414 0.888 0.980 0.020
#> SRR1274557 1 0.0376 0.891 0.996 0.004
#> SRR1274558 1 0.0000 0.890 1.000 0.000
#> SRR1274559 1 0.0000 0.890 1.000 0.000
#> SRR1274560 1 0.0000 0.890 1.000 0.000
#> SRR1274561 1 0.0000 0.890 1.000 0.000
#> SRR1274562 1 0.0000 0.890 1.000 0.000
#> SRR1274563 1 0.0000 0.890 1.000 0.000
#> SRR1274564 1 0.6148 0.815 0.848 0.152
#> SRR1274565 1 0.6148 0.815 0.848 0.152
#> SRR1274566 1 0.6148 0.815 0.848 0.152
#> SRR1274567 1 0.4939 0.844 0.892 0.108
#> SRR1274568 1 0.0000 0.890 1.000 0.000
#> SRR1274569 1 0.0000 0.890 1.000 0.000
#> SRR1274570 1 0.0000 0.890 1.000 0.000
#> SRR1274571 1 0.0000 0.890 1.000 0.000
#> SRR1274572 1 0.0000 0.890 1.000 0.000
#> SRR1274573 2 0.0000 0.911 0.000 1.000
#> SRR1274574 2 0.7219 0.708 0.200 0.800
#> SRR1274575 1 0.8763 0.657 0.704 0.296
#> SRR1274576 1 0.7815 0.743 0.768 0.232
#> SRR1274577 1 0.0000 0.890 1.000 0.000
#> SRR1274578 1 0.0000 0.890 1.000 0.000
#> SRR1274579 1 0.0000 0.890 1.000 0.000
#> SRR1274580 1 0.0000 0.890 1.000 0.000
#> SRR1274581 1 0.1184 0.889 0.984 0.016
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.5560 0.5393 0.000 0.700 0.300
#> SRR1274529 3 0.0424 0.7510 0.008 0.000 0.992
#> SRR1274530 3 0.0424 0.7510 0.008 0.000 0.992
#> SRR1274531 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274532 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274533 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274534 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274535 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274536 3 0.0000 0.7566 0.000 0.000 1.000
#> SRR1274537 3 0.6978 0.4173 0.032 0.336 0.632
#> SRR1274538 2 0.6309 0.0188 0.000 0.504 0.496
#> SRR1274539 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274540 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274543 3 0.4974 0.6557 0.000 0.236 0.764
#> SRR1274544 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274545 3 0.0000 0.7566 0.000 0.000 1.000
#> SRR1274546 2 0.1529 0.7467 0.000 0.960 0.040
#> SRR1274547 3 0.4842 0.6620 0.000 0.224 0.776
#> SRR1274548 2 0.6309 0.0188 0.000 0.504 0.496
#> SRR1274549 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274550 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274551 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274552 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274553 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274554 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274555 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274556 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274557 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274558 3 0.5760 0.7530 0.328 0.000 0.672
#> SRR1274559 3 0.5497 0.7937 0.292 0.000 0.708
#> SRR1274560 3 0.6286 0.4953 0.464 0.000 0.536
#> SRR1274561 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274562 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274563 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274564 3 0.0000 0.7566 0.000 0.000 1.000
#> SRR1274565 3 0.1031 0.7569 0.000 0.024 0.976
#> SRR1274566 3 0.0000 0.7566 0.000 0.000 1.000
#> SRR1274567 3 0.0000 0.7566 0.000 0.000 1.000
#> SRR1274568 3 0.5760 0.7530 0.328 0.000 0.672
#> SRR1274569 1 0.4931 0.5400 0.768 0.000 0.232
#> SRR1274570 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274571 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274572 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274573 2 0.0000 0.7655 0.000 1.000 0.000
#> SRR1274574 2 0.6008 0.5162 0.000 0.628 0.372
#> SRR1274575 3 0.0424 0.7510 0.008 0.000 0.992
#> SRR1274576 3 0.6225 0.1811 0.000 0.432 0.568
#> SRR1274577 3 0.4887 0.8475 0.228 0.000 0.772
#> SRR1274578 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274579 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274580 1 0.0424 0.9389 0.992 0.000 0.008
#> SRR1274581 3 0.4887 0.8475 0.228 0.000 0.772
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.5184 0.5373 0.000 0.672 0.304 0.024
#> SRR1274529 4 0.3801 1.0000 0.000 0.000 0.220 0.780
#> SRR1274530 4 0.3801 1.0000 0.000 0.000 0.220 0.780
#> SRR1274531 3 0.3852 0.5931 0.000 0.012 0.808 0.180
#> SRR1274532 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.2976 0.5825 0.000 0.008 0.872 0.120
#> SRR1274534 3 0.2976 0.5825 0.000 0.008 0.872 0.120
#> SRR1274535 3 0.2976 0.5825 0.000 0.008 0.872 0.120
#> SRR1274536 3 0.4933 0.1725 0.000 0.000 0.568 0.432
#> SRR1274537 3 0.5543 0.1501 0.000 0.360 0.612 0.028
#> SRR1274538 2 0.5691 0.2011 0.000 0.508 0.468 0.024
#> SRR1274539 3 0.3032 0.5841 0.000 0.008 0.868 0.124
#> SRR1274540 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274543 3 0.6277 0.1708 0.000 0.360 0.572 0.068
#> SRR1274544 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274545 3 0.6774 0.1736 0.000 0.120 0.568 0.312
#> SRR1274546 2 0.2760 0.7526 0.000 0.872 0.128 0.000
#> SRR1274547 3 0.7827 -0.0468 0.000 0.316 0.408 0.276
#> SRR1274548 3 0.5697 -0.2320 0.000 0.488 0.488 0.024
#> SRR1274549 3 0.5110 0.5097 0.012 0.000 0.636 0.352
#> SRR1274550 3 0.3400 0.6093 0.000 0.000 0.820 0.180
#> SRR1274551 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.2530 0.6216 0.000 0.000 0.888 0.112
#> SRR1274553 3 0.2530 0.6216 0.000 0.000 0.888 0.112
#> SRR1274554 3 0.2976 0.5825 0.000 0.008 0.872 0.120
#> SRR1274555 3 0.4304 0.5785 0.000 0.000 0.716 0.284
#> SRR1274556 3 0.4936 0.5511 0.012 0.000 0.672 0.316
#> SRR1274557 3 0.3837 0.6026 0.000 0.000 0.776 0.224
#> SRR1274558 3 0.1118 0.6261 0.036 0.000 0.964 0.000
#> SRR1274559 3 0.3156 0.6241 0.048 0.000 0.884 0.068
#> SRR1274560 3 0.2908 0.6247 0.064 0.000 0.896 0.040
#> SRR1274561 3 0.4910 0.5719 0.020 0.000 0.704 0.276
#> SRR1274562 3 0.4910 0.5719 0.020 0.000 0.704 0.276
#> SRR1274563 3 0.4222 0.5845 0.000 0.000 0.728 0.272
#> SRR1274564 3 0.4925 0.1732 0.000 0.000 0.572 0.428
#> SRR1274565 3 0.4464 0.5647 0.000 0.024 0.768 0.208
#> SRR1274566 3 0.4925 0.1732 0.000 0.000 0.572 0.428
#> SRR1274567 3 0.4941 0.1713 0.000 0.000 0.564 0.436
#> SRR1274568 3 0.1118 0.6261 0.036 0.000 0.964 0.000
#> SRR1274569 1 0.3123 0.7051 0.844 0.000 0.156 0.000
#> SRR1274570 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.8324 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.5458 0.5818 0.000 0.704 0.236 0.060
#> SRR1274575 4 0.3801 1.0000 0.000 0.000 0.220 0.780
#> SRR1274576 3 0.4903 0.4149 0.000 0.248 0.724 0.028
#> SRR1274577 3 0.0376 0.6270 0.004 0.000 0.992 0.004
#> SRR1274578 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.9578 1.000 0.000 0.000 0.000
#> SRR1274581 3 0.5130 0.5504 0.020 0.000 0.668 0.312
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.3048 0.69849 0.000 0.820 0.000 0.004 0.176
#> SRR1274529 4 0.1901 0.81876 0.000 0.012 0.056 0.928 0.004
#> SRR1274530 4 0.1901 0.81876 0.000 0.012 0.056 0.928 0.004
#> SRR1274531 5 0.6964 -0.00651 0.000 0.016 0.368 0.200 0.416
#> SRR1274532 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.1116 0.74969 0.000 0.004 0.028 0.004 0.964
#> SRR1274534 5 0.1412 0.74915 0.000 0.008 0.036 0.004 0.952
#> SRR1274535 5 0.0955 0.74831 0.000 0.000 0.028 0.004 0.968
#> SRR1274536 4 0.3210 0.81481 0.000 0.012 0.132 0.844 0.012
#> SRR1274537 5 0.4054 0.56977 0.000 0.224 0.000 0.028 0.748
#> SRR1274538 2 0.4990 0.29862 0.000 0.580 0.000 0.036 0.384
#> SRR1274539 5 0.1329 0.75001 0.000 0.008 0.032 0.004 0.956
#> SRR1274540 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 5 0.5990 -0.02982 0.000 0.436 0.024 0.056 0.484
#> SRR1274544 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 4 0.6151 0.68381 0.000 0.064 0.144 0.664 0.128
#> SRR1274546 2 0.1768 0.78817 0.000 0.924 0.000 0.004 0.072
#> SRR1274547 4 0.7542 0.48619 0.000 0.164 0.128 0.520 0.188
#> SRR1274548 2 0.4747 0.06671 0.000 0.500 0.000 0.016 0.484
#> SRR1274549 3 0.3895 0.32533 0.000 0.000 0.680 0.320 0.000
#> SRR1274550 3 0.4066 0.53621 0.000 0.000 0.672 0.004 0.324
#> SRR1274551 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.4367 0.40126 0.000 0.000 0.580 0.004 0.416
#> SRR1274553 3 0.4238 0.48609 0.000 0.000 0.628 0.004 0.368
#> SRR1274554 5 0.0955 0.74831 0.000 0.000 0.028 0.004 0.968
#> SRR1274555 3 0.1106 0.72875 0.000 0.000 0.964 0.024 0.012
#> SRR1274556 3 0.1671 0.70200 0.000 0.000 0.924 0.076 0.000
#> SRR1274557 3 0.2616 0.72344 0.000 0.000 0.880 0.020 0.100
#> SRR1274558 5 0.4574 0.60723 0.008 0.000 0.184 0.060 0.748
#> SRR1274559 3 0.6472 0.16297 0.048 0.000 0.504 0.068 0.380
#> SRR1274560 3 0.6548 0.19591 0.060 0.000 0.492 0.060 0.388
#> SRR1274561 3 0.1799 0.73566 0.020 0.000 0.940 0.012 0.028
#> SRR1274562 3 0.1074 0.73364 0.004 0.000 0.968 0.012 0.016
#> SRR1274563 3 0.1195 0.73586 0.000 0.000 0.960 0.012 0.028
#> SRR1274564 4 0.3018 0.82240 0.000 0.012 0.116 0.860 0.012
#> SRR1274565 4 0.6471 0.62239 0.000 0.036 0.180 0.604 0.180
#> SRR1274566 4 0.3018 0.82234 0.000 0.012 0.116 0.860 0.012
#> SRR1274567 4 0.3858 0.73043 0.000 0.008 0.224 0.760 0.008
#> SRR1274568 5 0.4391 0.62590 0.008 0.000 0.164 0.060 0.768
#> SRR1274569 1 0.0703 0.96999 0.976 0.000 0.024 0.000 0.000
#> SRR1274570 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.82745 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.7470 0.37960 0.000 0.524 0.112 0.200 0.164
#> SRR1274575 4 0.1901 0.81876 0.000 0.012 0.056 0.928 0.004
#> SRR1274576 5 0.3209 0.64755 0.000 0.180 0.000 0.008 0.812
#> SRR1274577 5 0.3264 0.64542 0.000 0.000 0.164 0.016 0.820
#> SRR1274578 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.99505 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.1764 0.70637 0.008 0.000 0.928 0.064 0.000
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.1970 0.741 0.000 0.900 0.000 0.000 0.008 0.092
#> SRR1274529 4 0.1204 0.743 0.000 0.056 0.000 0.944 0.000 0.000
#> SRR1274530 4 0.1204 0.743 0.000 0.056 0.000 0.944 0.000 0.000
#> SRR1274531 3 0.4901 0.685 0.000 0.052 0.712 0.068 0.168 0.000
#> SRR1274532 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274533 5 0.0713 0.832 0.000 0.028 0.000 0.000 0.972 0.000
#> SRR1274534 5 0.0713 0.832 0.000 0.028 0.000 0.000 0.972 0.000
#> SRR1274535 5 0.0713 0.832 0.000 0.028 0.000 0.000 0.972 0.000
#> SRR1274536 4 0.3647 0.536 0.000 0.000 0.360 0.640 0.000 0.000
#> SRR1274537 2 0.1858 0.813 0.000 0.904 0.000 0.000 0.092 0.004
#> SRR1274538 2 0.1806 0.815 0.000 0.908 0.000 0.000 0.088 0.004
#> SRR1274539 5 0.0713 0.832 0.000 0.028 0.000 0.000 0.972 0.000
#> SRR1274540 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274541 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274542 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274543 2 0.2218 0.807 0.000 0.884 0.000 0.012 0.104 0.000
#> SRR1274544 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274545 4 0.3290 0.609 0.000 0.252 0.004 0.744 0.000 0.000
#> SRR1274546 2 0.2883 0.575 0.000 0.788 0.000 0.000 0.000 0.212
#> SRR1274547 2 0.3405 0.549 0.000 0.724 0.004 0.272 0.000 0.000
#> SRR1274548 2 0.1806 0.815 0.000 0.908 0.000 0.000 0.088 0.004
#> SRR1274549 3 0.2151 0.812 0.000 0.000 0.904 0.072 0.016 0.008
#> SRR1274550 3 0.2219 0.821 0.000 0.000 0.864 0.000 0.136 0.000
#> SRR1274551 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274552 3 0.2416 0.814 0.000 0.000 0.844 0.000 0.156 0.000
#> SRR1274553 3 0.2378 0.817 0.000 0.000 0.848 0.000 0.152 0.000
#> SRR1274554 5 0.0713 0.832 0.000 0.028 0.000 0.000 0.972 0.000
#> SRR1274555 3 0.1847 0.841 0.000 0.008 0.928 0.008 0.048 0.008
#> SRR1274556 3 0.2036 0.815 0.000 0.000 0.912 0.064 0.016 0.008
#> SRR1274557 3 0.1668 0.842 0.000 0.008 0.928 0.004 0.060 0.000
#> SRR1274558 2 0.6532 0.304 0.000 0.532 0.060 0.004 0.224 0.180
#> SRR1274559 3 0.6435 0.553 0.020 0.092 0.596 0.004 0.076 0.212
#> SRR1274560 3 0.6598 0.551 0.032 0.092 0.592 0.004 0.076 0.204
#> SRR1274561 3 0.1908 0.832 0.000 0.020 0.924 0.000 0.012 0.044
#> SRR1274562 3 0.1225 0.837 0.000 0.000 0.952 0.000 0.012 0.036
#> SRR1274563 3 0.1333 0.841 0.000 0.008 0.944 0.000 0.048 0.000
#> SRR1274564 4 0.2562 0.754 0.000 0.000 0.172 0.828 0.000 0.000
#> SRR1274565 4 0.4951 0.687 0.000 0.088 0.116 0.724 0.072 0.000
#> SRR1274566 4 0.2854 0.736 0.000 0.000 0.208 0.792 0.000 0.000
#> SRR1274567 4 0.3774 0.434 0.000 0.000 0.408 0.592 0.000 0.000
#> SRR1274568 5 0.7425 0.325 0.004 0.128 0.268 0.004 0.420 0.176
#> SRR1274569 1 0.2263 0.887 0.900 0.004 0.060 0.000 0.000 0.036
#> SRR1274570 1 0.0000 0.982 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.982 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.982 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 6 0.2969 1.000 0.000 0.224 0.000 0.000 0.000 0.776
#> SRR1274574 2 0.2981 0.681 0.000 0.820 0.000 0.160 0.000 0.020
#> SRR1274575 4 0.1204 0.743 0.000 0.056 0.000 0.944 0.000 0.000
#> SRR1274576 2 0.1806 0.815 0.000 0.908 0.000 0.000 0.088 0.004
#> SRR1274577 5 0.6669 0.539 0.008 0.116 0.192 0.000 0.556 0.128
#> SRR1274578 1 0.0146 0.979 0.996 0.004 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.982 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.982 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.3324 0.781 0.000 0.024 0.840 0.088 0.000 0.048
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["MAD", "NMF"]
# you can also extract it by
# res = res_list["MAD:NMF"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'MAD' method.
#> Subgroups are detected by 'NMF' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.885 0.932 0.971 0.4912 0.508 0.508
#> 3 3 0.662 0.684 0.833 0.2903 0.778 0.596
#> 4 4 0.813 0.891 0.934 0.1585 0.783 0.493
#> 5 5 0.835 0.807 0.887 0.0476 0.963 0.864
#> 6 6 0.816 0.733 0.869 0.0396 0.954 0.812
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.971 0.000 1.000
#> SRR1274529 2 0.000 0.971 0.000 1.000
#> SRR1274530 2 0.000 0.971 0.000 1.000
#> SRR1274531 2 0.605 0.829 0.148 0.852
#> SRR1274532 2 0.000 0.971 0.000 1.000
#> SRR1274533 2 0.000 0.971 0.000 1.000
#> SRR1274534 2 0.000 0.971 0.000 1.000
#> SRR1274535 1 0.991 0.171 0.556 0.444
#> SRR1274536 2 0.625 0.821 0.156 0.844
#> SRR1274537 2 0.000 0.971 0.000 1.000
#> SRR1274538 2 0.000 0.971 0.000 1.000
#> SRR1274539 2 0.000 0.971 0.000 1.000
#> SRR1274540 2 0.000 0.971 0.000 1.000
#> SRR1274541 2 0.000 0.971 0.000 1.000
#> SRR1274542 2 0.000 0.971 0.000 1.000
#> SRR1274543 2 0.000 0.971 0.000 1.000
#> SRR1274544 2 0.000 0.971 0.000 1.000
#> SRR1274545 2 0.000 0.971 0.000 1.000
#> SRR1274546 2 0.000 0.971 0.000 1.000
#> SRR1274547 2 0.000 0.971 0.000 1.000
#> SRR1274548 2 0.000 0.971 0.000 1.000
#> SRR1274549 1 0.000 0.965 1.000 0.000
#> SRR1274550 1 0.000 0.965 1.000 0.000
#> SRR1274551 2 0.000 0.971 0.000 1.000
#> SRR1274552 1 0.000 0.965 1.000 0.000
#> SRR1274553 1 0.000 0.965 1.000 0.000
#> SRR1274554 2 0.760 0.729 0.220 0.780
#> SRR1274555 1 0.000 0.965 1.000 0.000
#> SRR1274556 1 0.000 0.965 1.000 0.000
#> SRR1274557 1 0.000 0.965 1.000 0.000
#> SRR1274558 2 0.000 0.971 0.000 1.000
#> SRR1274559 1 0.163 0.944 0.976 0.024
#> SRR1274560 1 0.000 0.965 1.000 0.000
#> SRR1274561 1 0.000 0.965 1.000 0.000
#> SRR1274562 1 0.000 0.965 1.000 0.000
#> SRR1274563 1 0.000 0.965 1.000 0.000
#> SRR1274564 2 0.000 0.971 0.000 1.000
#> SRR1274565 2 0.000 0.971 0.000 1.000
#> SRR1274566 2 0.000 0.971 0.000 1.000
#> SRR1274567 1 0.767 0.694 0.776 0.224
#> SRR1274568 2 0.327 0.922 0.060 0.940
#> SRR1274569 1 0.000 0.965 1.000 0.000
#> SRR1274570 1 0.000 0.965 1.000 0.000
#> SRR1274571 1 0.000 0.965 1.000 0.000
#> SRR1274572 1 0.000 0.965 1.000 0.000
#> SRR1274573 2 0.000 0.971 0.000 1.000
#> SRR1274574 2 0.000 0.971 0.000 1.000
#> SRR1274575 2 0.000 0.971 0.000 1.000
#> SRR1274576 2 0.000 0.971 0.000 1.000
#> SRR1274577 2 0.839 0.647 0.268 0.732
#> SRR1274578 1 0.000 0.965 1.000 0.000
#> SRR1274579 1 0.000 0.965 1.000 0.000
#> SRR1274580 1 0.000 0.965 1.000 0.000
#> SRR1274581 1 0.000 0.965 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274529 2 0.0000 0.6780 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.6780 0.000 1.000 0.000
#> SRR1274531 3 0.4842 0.6120 0.000 0.224 0.776
#> SRR1274532 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274533 2 0.9664 0.5978 0.332 0.444 0.224
#> SRR1274534 2 0.6473 0.8548 0.332 0.652 0.016
#> SRR1274535 3 0.5733 0.4877 0.324 0.000 0.676
#> SRR1274536 3 0.5810 0.5660 0.000 0.336 0.664
#> SRR1274537 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274538 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274539 3 0.9942 -0.2676 0.332 0.288 0.380
#> SRR1274540 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274541 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274542 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274543 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274544 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274545 2 0.0000 0.6780 0.000 1.000 0.000
#> SRR1274546 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274547 2 0.0892 0.6922 0.020 0.980 0.000
#> SRR1274548 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274549 3 0.5621 0.5798 0.000 0.308 0.692
#> SRR1274550 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274551 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274552 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274553 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274554 3 0.9452 0.0969 0.284 0.220 0.496
#> SRR1274555 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274556 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274557 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274558 2 0.6204 0.7956 0.424 0.576 0.000
#> SRR1274559 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274560 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274561 3 0.4842 0.2538 0.224 0.000 0.776
#> SRR1274562 1 0.6295 0.6183 0.528 0.000 0.472
#> SRR1274563 3 0.0000 0.6541 0.000 0.000 1.000
#> SRR1274564 2 0.1860 0.6231 0.000 0.948 0.052
#> SRR1274565 2 0.0000 0.6780 0.000 1.000 0.000
#> SRR1274566 2 0.3038 0.5491 0.000 0.896 0.104
#> SRR1274567 3 0.5948 0.5562 0.000 0.360 0.640
#> SRR1274568 1 0.6215 -0.6521 0.572 0.428 0.000
#> SRR1274569 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274570 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274571 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274572 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274573 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274574 2 0.5733 0.8635 0.324 0.676 0.000
#> SRR1274575 2 0.0000 0.6780 0.000 1.000 0.000
#> SRR1274576 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274577 2 0.5785 0.8662 0.332 0.668 0.000
#> SRR1274578 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274579 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274580 1 0.5785 0.8482 0.668 0.000 0.332
#> SRR1274581 3 0.3412 0.4874 0.124 0.000 0.876
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0592 0.912 0.000 0.984 0.000 0.016
#> SRR1274529 4 0.1716 0.952 0.000 0.064 0.000 0.936
#> SRR1274530 4 0.1716 0.952 0.000 0.064 0.000 0.936
#> SRR1274531 3 0.1284 0.916 0.000 0.012 0.964 0.024
#> SRR1274532 2 0.1389 0.911 0.000 0.952 0.000 0.048
#> SRR1274533 2 0.3494 0.780 0.000 0.824 0.172 0.004
#> SRR1274534 2 0.1902 0.876 0.000 0.932 0.064 0.004
#> SRR1274535 3 0.4761 0.474 0.000 0.332 0.664 0.004
#> SRR1274536 4 0.1557 0.884 0.000 0.000 0.056 0.944
#> SRR1274537 2 0.0188 0.910 0.000 0.996 0.000 0.004
#> SRR1274538 2 0.1792 0.903 0.000 0.932 0.000 0.068
#> SRR1274539 2 0.2197 0.867 0.000 0.916 0.080 0.004
#> SRR1274540 2 0.1716 0.905 0.000 0.936 0.000 0.064
#> SRR1274541 2 0.1211 0.912 0.000 0.960 0.000 0.040
#> SRR1274542 2 0.1389 0.911 0.000 0.952 0.000 0.048
#> SRR1274543 2 0.0817 0.913 0.000 0.976 0.000 0.024
#> SRR1274544 2 0.1474 0.910 0.000 0.948 0.000 0.052
#> SRR1274545 4 0.1792 0.949 0.000 0.068 0.000 0.932
#> SRR1274546 2 0.0921 0.913 0.000 0.972 0.000 0.028
#> SRR1274547 4 0.2011 0.939 0.000 0.080 0.000 0.920
#> SRR1274548 2 0.0592 0.912 0.000 0.984 0.000 0.016
#> SRR1274549 3 0.1716 0.908 0.000 0.000 0.936 0.064
#> SRR1274550 3 0.0000 0.913 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.1867 0.900 0.000 0.928 0.000 0.072
#> SRR1274552 3 0.1004 0.902 0.000 0.024 0.972 0.004
#> SRR1274553 3 0.1004 0.902 0.000 0.024 0.972 0.004
#> SRR1274554 2 0.4699 0.546 0.000 0.676 0.320 0.004
#> SRR1274555 3 0.1389 0.916 0.000 0.000 0.952 0.048
#> SRR1274556 3 0.1557 0.913 0.000 0.000 0.944 0.056
#> SRR1274557 3 0.0592 0.917 0.000 0.000 0.984 0.016
#> SRR1274558 2 0.3583 0.795 0.180 0.816 0.000 0.004
#> SRR1274559 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274560 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.4514 0.775 0.800 0.000 0.136 0.064
#> SRR1274562 1 0.2610 0.885 0.900 0.000 0.088 0.012
#> SRR1274563 3 0.1389 0.916 0.000 0.000 0.952 0.048
#> SRR1274564 4 0.0336 0.934 0.000 0.008 0.000 0.992
#> SRR1274565 4 0.1716 0.951 0.000 0.064 0.000 0.936
#> SRR1274566 4 0.0188 0.931 0.000 0.004 0.000 0.996
#> SRR1274567 4 0.1302 0.896 0.000 0.000 0.044 0.956
#> SRR1274568 2 0.3266 0.805 0.168 0.832 0.000 0.000
#> SRR1274569 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.1557 0.908 0.000 0.944 0.000 0.056
#> SRR1274574 2 0.4331 0.622 0.000 0.712 0.000 0.288
#> SRR1274575 4 0.1716 0.952 0.000 0.064 0.000 0.936
#> SRR1274576 2 0.0524 0.905 0.000 0.988 0.008 0.004
#> SRR1274577 2 0.1743 0.881 0.000 0.940 0.056 0.004
#> SRR1274578 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.970 1.000 0.000 0.000 0.000
#> SRR1274581 3 0.4130 0.822 0.108 0.000 0.828 0.064
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0162 0.895 0.000 0.996 0.000 0.004 0.000
#> SRR1274529 4 0.0290 0.939 0.000 0.008 0.000 0.992 0.000
#> SRR1274530 4 0.0290 0.939 0.000 0.008 0.000 0.992 0.000
#> SRR1274531 3 0.4746 0.840 0.000 0.000 0.600 0.024 0.376
#> SRR1274532 2 0.0290 0.895 0.000 0.992 0.000 0.008 0.000
#> SRR1274533 5 0.3752 0.465 0.000 0.292 0.000 0.000 0.708
#> SRR1274534 2 0.4300 0.100 0.000 0.524 0.000 0.000 0.476
#> SRR1274535 5 0.1270 0.745 0.000 0.052 0.000 0.000 0.948
#> SRR1274536 4 0.0404 0.932 0.000 0.000 0.000 0.988 0.012
#> SRR1274537 2 0.0000 0.893 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0510 0.892 0.000 0.984 0.000 0.016 0.000
#> SRR1274539 2 0.4306 0.041 0.000 0.508 0.000 0.000 0.492
#> SRR1274540 2 0.0510 0.892 0.000 0.984 0.000 0.016 0.000
#> SRR1274541 2 0.0290 0.895 0.000 0.992 0.000 0.008 0.000
#> SRR1274542 2 0.0290 0.895 0.000 0.992 0.000 0.008 0.000
#> SRR1274543 2 0.0162 0.895 0.000 0.996 0.000 0.004 0.000
#> SRR1274544 2 0.0290 0.895 0.000 0.992 0.000 0.008 0.000
#> SRR1274545 4 0.1270 0.913 0.000 0.052 0.000 0.948 0.000
#> SRR1274546 2 0.0162 0.895 0.000 0.996 0.000 0.004 0.000
#> SRR1274547 4 0.1197 0.917 0.000 0.048 0.000 0.952 0.000
#> SRR1274548 2 0.0000 0.893 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 3 0.5538 0.788 0.000 0.000 0.504 0.068 0.428
#> SRR1274550 5 0.0963 0.720 0.000 0.000 0.036 0.000 0.964
#> SRR1274551 2 0.0404 0.894 0.000 0.988 0.000 0.012 0.000
#> SRR1274552 5 0.0162 0.763 0.000 0.000 0.004 0.000 0.996
#> SRR1274553 5 0.0290 0.761 0.000 0.000 0.008 0.000 0.992
#> SRR1274554 2 0.5372 0.521 0.000 0.668 0.152 0.000 0.180
#> SRR1274555 3 0.4383 0.829 0.000 0.000 0.572 0.004 0.424
#> SRR1274556 3 0.4029 0.828 0.000 0.000 0.680 0.004 0.316
#> SRR1274557 3 0.3983 0.836 0.000 0.000 0.660 0.000 0.340
#> SRR1274558 2 0.2233 0.822 0.104 0.892 0.000 0.004 0.000
#> SRR1274559 1 0.0703 0.911 0.976 0.000 0.024 0.000 0.000
#> SRR1274560 1 0.0510 0.913 0.984 0.000 0.016 0.000 0.000
#> SRR1274561 1 0.5562 0.477 0.624 0.000 0.300 0.056 0.020
#> SRR1274562 1 0.3476 0.744 0.804 0.000 0.176 0.000 0.020
#> SRR1274563 3 0.4425 0.806 0.004 0.000 0.544 0.000 0.452
#> SRR1274564 4 0.0000 0.936 0.000 0.000 0.000 1.000 0.000
#> SRR1274565 4 0.5444 0.579 0.000 0.180 0.160 0.660 0.000
#> SRR1274566 4 0.0404 0.935 0.000 0.000 0.012 0.988 0.000
#> SRR1274567 4 0.1106 0.922 0.000 0.000 0.024 0.964 0.012
#> SRR1274568 2 0.3090 0.800 0.104 0.856 0.000 0.000 0.040
#> SRR1274569 1 0.0000 0.914 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0162 0.914 0.996 0.000 0.004 0.000 0.000
#> SRR1274571 1 0.0162 0.914 0.996 0.000 0.004 0.000 0.000
#> SRR1274572 1 0.0404 0.913 0.988 0.000 0.012 0.000 0.000
#> SRR1274573 2 0.0290 0.895 0.000 0.992 0.000 0.008 0.000
#> SRR1274574 2 0.1478 0.858 0.000 0.936 0.000 0.064 0.000
#> SRR1274575 4 0.0290 0.939 0.000 0.008 0.000 0.992 0.000
#> SRR1274576 2 0.0703 0.881 0.000 0.976 0.000 0.000 0.024
#> SRR1274577 2 0.4489 0.708 0.004 0.764 0.092 0.000 0.140
#> SRR1274578 1 0.1792 0.892 0.916 0.000 0.084 0.000 0.000
#> SRR1274579 1 0.1792 0.892 0.916 0.000 0.084 0.000 0.000
#> SRR1274580 1 0.1671 0.896 0.924 0.000 0.076 0.000 0.000
#> SRR1274581 3 0.3080 0.536 0.008 0.000 0.844 0.008 0.140
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0260 0.897 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274529 4 0.0363 0.886 0.000 0.012 0.000 0.988 0.000 0.000
#> SRR1274530 4 0.0146 0.888 0.000 0.004 0.000 0.996 0.000 0.000
#> SRR1274531 3 0.2546 0.831 0.000 0.004 0.896 0.040 0.040 0.020
#> SRR1274532 2 0.0146 0.901 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274533 5 0.3093 0.781 0.000 0.076 0.060 0.000 0.852 0.012
#> SRR1274534 5 0.3900 0.516 0.000 0.232 0.000 0.000 0.728 0.040
#> SRR1274535 5 0.2308 0.812 0.000 0.008 0.108 0.000 0.880 0.004
#> SRR1274536 4 0.0692 0.882 0.000 0.000 0.004 0.976 0.000 0.020
#> SRR1274537 2 0.0260 0.897 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274538 2 0.0458 0.892 0.000 0.984 0.000 0.016 0.000 0.000
#> SRR1274539 5 0.4112 0.589 0.000 0.200 0.016 0.000 0.744 0.040
#> SRR1274540 2 0.0260 0.899 0.000 0.992 0.000 0.008 0.000 0.000
#> SRR1274541 2 0.0146 0.901 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274542 2 0.0146 0.901 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274543 2 0.0260 0.897 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274544 2 0.0146 0.901 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274545 4 0.1765 0.825 0.000 0.096 0.000 0.904 0.000 0.000
#> SRR1274546 2 0.0000 0.900 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 4 0.2003 0.800 0.000 0.116 0.000 0.884 0.000 0.000
#> SRR1274548 2 0.0520 0.891 0.000 0.984 0.000 0.000 0.008 0.008
#> SRR1274549 3 0.3864 0.757 0.000 0.000 0.796 0.128 0.044 0.032
#> SRR1274550 5 0.2631 0.762 0.000 0.000 0.180 0.000 0.820 0.000
#> SRR1274551 2 0.0260 0.899 0.000 0.992 0.000 0.008 0.000 0.000
#> SRR1274552 5 0.2048 0.811 0.000 0.000 0.120 0.000 0.880 0.000
#> SRR1274553 5 0.2003 0.812 0.000 0.000 0.116 0.000 0.884 0.000
#> SRR1274554 2 0.6721 -0.381 0.000 0.484 0.220 0.000 0.068 0.228
#> SRR1274555 3 0.1610 0.828 0.000 0.000 0.916 0.000 0.084 0.000
#> SRR1274556 3 0.0713 0.830 0.000 0.000 0.972 0.000 0.000 0.028
#> SRR1274557 3 0.0363 0.838 0.000 0.000 0.988 0.000 0.012 0.000
#> SRR1274558 2 0.2491 0.651 0.164 0.836 0.000 0.000 0.000 0.000
#> SRR1274559 1 0.1655 0.750 0.940 0.032 0.000 0.012 0.004 0.012
#> SRR1274560 1 0.0508 0.774 0.984 0.000 0.000 0.000 0.004 0.012
#> SRR1274561 1 0.4628 0.426 0.648 0.000 0.308 0.016 0.012 0.016
#> SRR1274562 1 0.3217 0.610 0.768 0.000 0.224 0.000 0.008 0.000
#> SRR1274563 3 0.2165 0.813 0.008 0.000 0.884 0.000 0.108 0.000
#> SRR1274564 4 0.0000 0.887 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274565 4 0.7150 0.185 0.000 0.164 0.148 0.492 0.008 0.188
#> SRR1274566 4 0.2002 0.856 0.000 0.000 0.012 0.908 0.004 0.076
#> SRR1274567 4 0.1844 0.864 0.000 0.000 0.024 0.924 0.004 0.048
#> SRR1274568 2 0.4056 0.585 0.028 0.780 0.000 0.000 0.056 0.136
#> SRR1274569 1 0.0146 0.776 0.996 0.000 0.000 0.000 0.000 0.004
#> SRR1274570 1 0.0000 0.776 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.776 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.1444 0.757 0.928 0.000 0.000 0.000 0.000 0.072
#> SRR1274573 2 0.0146 0.901 0.000 0.996 0.000 0.004 0.000 0.000
#> SRR1274574 2 0.0937 0.865 0.000 0.960 0.000 0.040 0.000 0.000
#> SRR1274575 4 0.0146 0.888 0.000 0.004 0.000 0.996 0.000 0.000
#> SRR1274576 2 0.2136 0.793 0.000 0.904 0.000 0.000 0.048 0.048
#> SRR1274577 6 0.5598 0.000 0.000 0.300 0.008 0.000 0.140 0.552
#> SRR1274578 1 0.4169 0.423 0.532 0.000 0.000 0.000 0.012 0.456
#> SRR1274579 1 0.3823 0.470 0.564 0.000 0.000 0.000 0.000 0.436
#> SRR1274580 1 0.3706 0.534 0.620 0.000 0.000 0.000 0.000 0.380
#> SRR1274581 3 0.5391 0.372 0.000 0.000 0.456 0.000 0.112 0.432
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "hclust"]
# you can also extract it by
# res = res_list["ATC:hclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'hclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.961 0.983 0.4377 0.575 0.575
#> 3 3 0.678 0.831 0.840 0.2002 0.980 0.966
#> 4 4 0.704 0.670 0.846 0.2523 0.754 0.557
#> 5 5 0.663 0.705 0.846 0.0175 0.904 0.730
#> 6 6 0.693 0.568 0.825 0.0614 0.881 0.652
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.975 0.000 1.000
#> SRR1274529 2 0.0000 0.975 0.000 1.000
#> SRR1274530 2 0.0000 0.975 0.000 1.000
#> SRR1274531 2 0.4161 0.899 0.084 0.916
#> SRR1274532 2 0.0000 0.975 0.000 1.000
#> SRR1274533 2 0.0000 0.975 0.000 1.000
#> SRR1274534 2 0.0000 0.975 0.000 1.000
#> SRR1274535 2 0.0672 0.969 0.008 0.992
#> SRR1274536 2 0.0376 0.972 0.004 0.996
#> SRR1274537 2 0.0000 0.975 0.000 1.000
#> SRR1274538 2 0.0000 0.975 0.000 1.000
#> SRR1274539 2 0.0000 0.975 0.000 1.000
#> SRR1274540 2 0.0000 0.975 0.000 1.000
#> SRR1274541 2 0.0000 0.975 0.000 1.000
#> SRR1274542 2 0.0000 0.975 0.000 1.000
#> SRR1274543 2 0.0000 0.975 0.000 1.000
#> SRR1274544 2 0.0000 0.975 0.000 1.000
#> SRR1274545 2 0.0000 0.975 0.000 1.000
#> SRR1274546 2 0.0000 0.975 0.000 1.000
#> SRR1274547 2 0.0000 0.975 0.000 1.000
#> SRR1274548 2 0.0000 0.975 0.000 1.000
#> SRR1274549 2 0.5629 0.847 0.132 0.868
#> SRR1274550 1 0.0000 1.000 1.000 0.000
#> SRR1274551 2 0.0000 0.975 0.000 1.000
#> SRR1274552 1 0.0000 1.000 1.000 0.000
#> SRR1274553 1 0.0000 1.000 1.000 0.000
#> SRR1274554 2 0.0000 0.975 0.000 1.000
#> SRR1274555 1 0.0000 1.000 1.000 0.000
#> SRR1274556 1 0.0000 1.000 1.000 0.000
#> SRR1274557 1 0.0000 1.000 1.000 0.000
#> SRR1274558 2 0.0000 0.975 0.000 1.000
#> SRR1274559 2 0.0000 0.975 0.000 1.000
#> SRR1274560 2 0.9129 0.539 0.328 0.672
#> SRR1274561 2 0.9427 0.472 0.360 0.640
#> SRR1274562 1 0.0000 1.000 1.000 0.000
#> SRR1274563 1 0.0000 1.000 1.000 0.000
#> SRR1274564 2 0.0376 0.972 0.004 0.996
#> SRR1274565 2 0.0000 0.975 0.000 1.000
#> SRR1274566 2 0.0376 0.972 0.004 0.996
#> SRR1274567 2 0.0376 0.972 0.004 0.996
#> SRR1274568 2 0.0000 0.975 0.000 1.000
#> SRR1274569 1 0.0000 1.000 1.000 0.000
#> SRR1274570 1 0.0000 1.000 1.000 0.000
#> SRR1274571 1 0.0000 1.000 1.000 0.000
#> SRR1274572 1 0.0000 1.000 1.000 0.000
#> SRR1274573 2 0.0000 0.975 0.000 1.000
#> SRR1274574 2 0.0000 0.975 0.000 1.000
#> SRR1274575 2 0.0000 0.975 0.000 1.000
#> SRR1274576 2 0.0000 0.975 0.000 1.000
#> SRR1274577 2 0.0000 0.975 0.000 1.000
#> SRR1274578 1 0.0000 1.000 1.000 0.000
#> SRR1274579 1 0.0000 1.000 1.000 0.000
#> SRR1274580 1 0.0000 1.000 1.000 0.000
#> SRR1274581 1 0.0000 1.000 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274529 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274530 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274531 2 0.464 0.749 0.084 0.856 0.060
#> SRR1274532 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274533 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274534 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274535 2 0.249 0.812 0.008 0.932 0.060
#> SRR1274536 2 0.230 0.814 0.004 0.936 0.060
#> SRR1274537 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274538 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274539 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274540 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274541 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274542 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274543 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274544 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274545 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274546 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274547 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274548 2 0.216 0.837 0.000 0.936 0.064
#> SRR1274549 2 0.554 0.697 0.132 0.808 0.060
#> SRR1274550 1 0.348 0.833 0.872 0.000 0.128
#> SRR1274551 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274552 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274553 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274554 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274555 1 0.348 0.833 0.872 0.000 0.128
#> SRR1274556 3 0.540 0.987 0.280 0.000 0.720
#> SRR1274557 1 0.175 0.902 0.952 0.000 0.048
#> SRR1274558 2 0.216 0.837 0.000 0.936 0.064
#> SRR1274559 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274560 2 0.760 0.338 0.328 0.612 0.060
#> SRR1274561 2 0.776 0.264 0.360 0.580 0.060
#> SRR1274562 1 0.348 0.833 0.872 0.000 0.128
#> SRR1274563 1 0.348 0.833 0.872 0.000 0.128
#> SRR1274564 2 0.230 0.814 0.004 0.936 0.060
#> SRR1274565 2 0.000 0.828 0.000 1.000 0.000
#> SRR1274566 2 0.230 0.814 0.004 0.936 0.060
#> SRR1274567 2 0.230 0.814 0.004 0.936 0.060
#> SRR1274568 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274569 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274570 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274571 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274572 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274573 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274574 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274575 2 0.470 0.841 0.000 0.788 0.212
#> SRR1274576 2 0.216 0.837 0.000 0.936 0.064
#> SRR1274577 2 0.207 0.816 0.000 0.940 0.060
#> SRR1274578 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274579 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274580 1 0.000 0.932 1.000 0.000 0.000
#> SRR1274581 3 0.533 0.988 0.272 0.000 0.728
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274531 4 0.4352 0.57730 0.080 0.104 0.000 0.816
#> SRR1274532 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274534 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274535 4 0.5257 0.51922 0.008 0.444 0.000 0.548
#> SRR1274536 4 0.1637 0.56985 0.000 0.060 0.000 0.940
#> SRR1274537 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274539 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274540 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.4761 0.06106 0.000 0.628 0.000 0.372
#> SRR1274549 4 0.2760 0.44026 0.128 0.000 0.000 0.872
#> SRR1274550 1 0.4855 0.53258 0.600 0.000 0.400 0.000
#> SRR1274551 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274552 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274553 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274554 4 0.4972 0.50697 0.000 0.456 0.000 0.544
#> SRR1274555 1 0.4855 0.53258 0.600 0.000 0.400 0.000
#> SRR1274556 3 0.0336 0.94461 0.008 0.000 0.992 0.000
#> SRR1274557 1 0.3444 0.73856 0.816 0.000 0.184 0.000
#> SRR1274558 2 0.4804 0.00817 0.000 0.616 0.000 0.384
#> SRR1274559 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274560 4 0.4543 0.35332 0.324 0.000 0.000 0.676
#> SRR1274561 4 0.4697 0.31070 0.356 0.000 0.000 0.644
#> SRR1274562 1 0.4855 0.53258 0.600 0.000 0.400 0.000
#> SRR1274563 1 0.4855 0.53258 0.600 0.000 0.400 0.000
#> SRR1274564 4 0.1637 0.56985 0.000 0.060 0.000 0.940
#> SRR1274565 2 0.4972 -0.29436 0.000 0.544 0.000 0.456
#> SRR1274566 4 0.1637 0.56985 0.000 0.060 0.000 0.940
#> SRR1274567 4 0.1637 0.56985 0.000 0.060 0.000 0.940
#> SRR1274568 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274569 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0000 0.88426 0.000 1.000 0.000 0.000
#> SRR1274576 2 0.4776 0.04450 0.000 0.624 0.000 0.376
#> SRR1274577 4 0.4967 0.51664 0.000 0.452 0.000 0.548
#> SRR1274578 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.84386 1.000 0.000 0.000 0.000
#> SRR1274581 3 0.1637 0.94575 0.000 0.000 0.940 0.060
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274529 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274530 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274531 4 0.5570 0.493 0.060 0.012 0.296 0.628 0.004
#> SRR1274532 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274533 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274534 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274535 4 0.0324 0.644 0.004 0.004 0.000 0.992 0.000
#> SRR1274536 4 0.4161 0.455 0.000 0.000 0.392 0.608 0.000
#> SRR1274537 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274538 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274539 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274540 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274541 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274542 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274543 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274544 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274545 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274546 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274547 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274548 4 0.2891 0.296 0.000 0.176 0.000 0.824 0.000
#> SRR1274549 4 0.6429 0.333 0.108 0.012 0.396 0.480 0.004
#> SRR1274550 1 0.4341 0.516 0.592 0.004 0.404 0.000 0.000
#> SRR1274551 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274552 1 0.0290 0.836 0.992 0.000 0.008 0.000 0.000
#> SRR1274553 1 0.0290 0.836 0.992 0.000 0.008 0.000 0.000
#> SRR1274554 4 0.0162 0.640 0.000 0.004 0.000 0.996 0.000
#> SRR1274555 1 0.4341 0.516 0.592 0.004 0.404 0.000 0.000
#> SRR1274556 3 0.4171 0.000 0.000 0.396 0.604 0.000 0.000
#> SRR1274557 1 0.3160 0.728 0.808 0.004 0.188 0.000 0.000
#> SRR1274558 4 0.2773 0.334 0.000 0.164 0.000 0.836 0.000
#> SRR1274559 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274560 4 0.6735 0.270 0.304 0.012 0.172 0.508 0.004
#> SRR1274561 4 0.6821 0.231 0.336 0.012 0.172 0.476 0.004
#> SRR1274562 1 0.4321 0.521 0.600 0.004 0.396 0.000 0.000
#> SRR1274563 1 0.4341 0.516 0.592 0.004 0.404 0.000 0.000
#> SRR1274564 4 0.4161 0.455 0.000 0.000 0.392 0.608 0.000
#> SRR1274565 4 0.1908 0.504 0.000 0.092 0.000 0.908 0.000
#> SRR1274566 4 0.4161 0.455 0.000 0.000 0.392 0.608 0.000
#> SRR1274567 4 0.4161 0.455 0.000 0.000 0.392 0.608 0.000
#> SRR1274568 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274569 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274574 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274575 2 0.4201 1.000 0.000 0.592 0.000 0.408 0.000
#> SRR1274576 4 0.2852 0.310 0.000 0.172 0.000 0.828 0.000
#> SRR1274577 4 0.0000 0.644 0.000 0.000 0.000 1.000 0.000
#> SRR1274578 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.838 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 5 0.0162 0.000 0.000 0.000 0.004 0.000 0.996
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274529 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274530 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274531 4 0.5195 -0.0463 0.000 0.100 0.000 0.540 0.360 0
#> SRR1274532 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274533 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274534 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274535 4 0.4353 0.5377 0.000 0.384 0.000 0.588 0.028 0
#> SRR1274536 4 0.2053 0.2320 0.000 0.004 0.000 0.888 0.108 0
#> SRR1274537 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274538 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274539 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274540 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274541 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274542 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274543 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274544 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274545 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274546 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274547 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274548 2 0.3823 -0.1232 0.000 0.564 0.000 0.436 0.000 0
#> SRR1274549 4 0.3782 -0.2705 0.000 0.000 0.000 0.588 0.412 0
#> SRR1274550 3 0.0363 0.6719 0.012 0.000 0.988 0.000 0.000 0
#> SRR1274551 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274552 3 0.4084 0.5237 0.400 0.000 0.588 0.000 0.012 0
#> SRR1274553 3 0.4084 0.5237 0.400 0.000 0.588 0.000 0.012 0
#> SRR1274554 4 0.3737 0.5381 0.000 0.392 0.000 0.608 0.000 0
#> SRR1274555 3 0.0363 0.6719 0.012 0.000 0.988 0.000 0.000 0
#> SRR1274556 5 0.3782 -0.3483 0.000 0.000 0.412 0.000 0.588 0
#> SRR1274557 3 0.3190 0.6715 0.220 0.000 0.772 0.000 0.008 0
#> SRR1274558 2 0.3838 -0.1686 0.000 0.552 0.000 0.448 0.000 0
#> SRR1274559 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274560 4 0.5993 -0.4755 0.232 0.000 0.000 0.392 0.376 0
#> SRR1274561 5 0.5986 -0.0397 0.232 0.000 0.000 0.360 0.408 0
#> SRR1274562 1 0.3789 0.2750 0.584 0.000 0.416 0.000 0.000 0
#> SRR1274563 3 0.0363 0.6719 0.012 0.000 0.988 0.000 0.000 0
#> SRR1274564 4 0.2053 0.2320 0.000 0.004 0.000 0.888 0.108 0
#> SRR1274565 4 0.3864 0.3295 0.000 0.480 0.000 0.520 0.000 0
#> SRR1274566 4 0.2053 0.2320 0.000 0.004 0.000 0.888 0.108 0
#> SRR1274567 4 0.2053 0.2320 0.000 0.004 0.000 0.888 0.108 0
#> SRR1274568 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274569 1 0.0622 0.9109 0.980 0.000 0.012 0.000 0.008 0
#> SRR1274570 1 0.0622 0.9109 0.980 0.000 0.012 0.000 0.008 0
#> SRR1274571 1 0.0622 0.9109 0.980 0.000 0.012 0.000 0.008 0
#> SRR1274572 1 0.0458 0.9097 0.984 0.000 0.016 0.000 0.000 0
#> SRR1274573 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274574 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274575 2 0.0000 0.9037 0.000 1.000 0.000 0.000 0.000 0
#> SRR1274576 2 0.3828 -0.1379 0.000 0.560 0.000 0.440 0.000 0
#> SRR1274577 4 0.3717 0.5507 0.000 0.384 0.000 0.616 0.000 0
#> SRR1274578 1 0.0000 0.9101 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274579 1 0.0000 0.9101 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274580 1 0.0000 0.9101 1.000 0.000 0.000 0.000 0.000 0
#> SRR1274581 6 0.0000 0.0000 0.000 0.000 0.000 0.000 0.000 1
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "kmeans"]
# you can also extract it by
# res = res_list["ATC:kmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'kmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.975 0.989 0.4740 0.525 0.525
#> 3 3 0.829 0.889 0.926 0.3837 0.737 0.529
#> 4 4 0.722 0.699 0.822 0.1019 0.960 0.883
#> 5 5 0.724 0.676 0.793 0.0640 0.904 0.703
#> 6 6 0.738 0.683 0.783 0.0524 0.891 0.574
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.992 0.000 1.000
#> SRR1274529 2 0.000 0.992 0.000 1.000
#> SRR1274530 2 0.000 0.992 0.000 1.000
#> SRR1274531 2 0.343 0.936 0.064 0.936
#> SRR1274532 2 0.000 0.992 0.000 1.000
#> SRR1274533 2 0.000 0.992 0.000 1.000
#> SRR1274534 2 0.000 0.992 0.000 1.000
#> SRR1274535 1 0.000 0.982 1.000 0.000
#> SRR1274536 2 0.000 0.992 0.000 1.000
#> SRR1274537 2 0.000 0.992 0.000 1.000
#> SRR1274538 2 0.000 0.992 0.000 1.000
#> SRR1274539 2 0.343 0.936 0.064 0.936
#> SRR1274540 2 0.000 0.992 0.000 1.000
#> SRR1274541 2 0.000 0.992 0.000 1.000
#> SRR1274542 2 0.000 0.992 0.000 1.000
#> SRR1274543 2 0.000 0.992 0.000 1.000
#> SRR1274544 2 0.000 0.992 0.000 1.000
#> SRR1274545 2 0.000 0.992 0.000 1.000
#> SRR1274546 2 0.000 0.992 0.000 1.000
#> SRR1274547 2 0.000 0.992 0.000 1.000
#> SRR1274548 2 0.000 0.992 0.000 1.000
#> SRR1274549 1 0.000 0.982 1.000 0.000
#> SRR1274550 1 0.000 0.982 1.000 0.000
#> SRR1274551 2 0.000 0.992 0.000 1.000
#> SRR1274552 1 0.000 0.982 1.000 0.000
#> SRR1274553 1 0.000 0.982 1.000 0.000
#> SRR1274554 2 0.000 0.992 0.000 1.000
#> SRR1274555 1 0.000 0.982 1.000 0.000
#> SRR1274556 1 0.000 0.982 1.000 0.000
#> SRR1274557 1 0.000 0.982 1.000 0.000
#> SRR1274558 2 0.000 0.992 0.000 1.000
#> SRR1274559 2 0.343 0.936 0.064 0.936
#> SRR1274560 1 0.913 0.501 0.672 0.328
#> SRR1274561 1 0.000 0.982 1.000 0.000
#> SRR1274562 1 0.000 0.982 1.000 0.000
#> SRR1274563 1 0.000 0.982 1.000 0.000
#> SRR1274564 2 0.000 0.992 0.000 1.000
#> SRR1274565 2 0.000 0.992 0.000 1.000
#> SRR1274566 2 0.000 0.992 0.000 1.000
#> SRR1274567 2 0.000 0.992 0.000 1.000
#> SRR1274568 2 0.000 0.992 0.000 1.000
#> SRR1274569 1 0.000 0.982 1.000 0.000
#> SRR1274570 1 0.000 0.982 1.000 0.000
#> SRR1274571 1 0.000 0.982 1.000 0.000
#> SRR1274572 1 0.000 0.982 1.000 0.000
#> SRR1274573 2 0.000 0.992 0.000 1.000
#> SRR1274574 2 0.000 0.992 0.000 1.000
#> SRR1274575 2 0.000 0.992 0.000 1.000
#> SRR1274576 2 0.000 0.992 0.000 1.000
#> SRR1274577 2 0.343 0.936 0.064 0.936
#> SRR1274578 1 0.000 0.982 1.000 0.000
#> SRR1274579 1 0.000 0.982 1.000 0.000
#> SRR1274580 1 0.000 0.982 1.000 0.000
#> SRR1274581 1 0.000 0.982 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274529 2 0.0592 0.966 0.000 0.988 0.012
#> SRR1274530 2 0.0592 0.966 0.000 0.988 0.012
#> SRR1274531 3 0.2066 0.888 0.000 0.060 0.940
#> SRR1274532 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274533 3 0.2796 0.890 0.000 0.092 0.908
#> SRR1274534 3 0.2356 0.890 0.000 0.072 0.928
#> SRR1274535 3 0.2050 0.854 0.028 0.020 0.952
#> SRR1274536 3 0.2878 0.885 0.000 0.096 0.904
#> SRR1274537 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274539 3 0.2356 0.890 0.000 0.072 0.928
#> SRR1274540 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274545 2 0.0592 0.966 0.000 0.988 0.012
#> SRR1274546 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274548 2 0.2066 0.918 0.000 0.940 0.060
#> SRR1274549 3 0.3192 0.767 0.112 0.000 0.888
#> SRR1274550 1 0.1753 0.917 0.952 0.000 0.048
#> SRR1274551 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274552 1 0.4399 0.877 0.812 0.000 0.188
#> SRR1274553 1 0.3686 0.915 0.860 0.000 0.140
#> SRR1274554 3 0.3038 0.886 0.000 0.104 0.896
#> SRR1274555 1 0.1753 0.917 0.952 0.000 0.048
#> SRR1274556 1 0.2066 0.913 0.940 0.000 0.060
#> SRR1274557 1 0.1753 0.917 0.952 0.000 0.048
#> SRR1274558 2 0.5760 0.440 0.000 0.672 0.328
#> SRR1274559 3 0.2066 0.888 0.000 0.060 0.940
#> SRR1274560 3 0.1529 0.833 0.040 0.000 0.960
#> SRR1274561 3 0.5859 0.295 0.344 0.000 0.656
#> SRR1274562 1 0.0000 0.912 1.000 0.000 0.000
#> SRR1274563 1 0.1753 0.917 0.952 0.000 0.048
#> SRR1274564 3 0.6286 0.196 0.000 0.464 0.536
#> SRR1274565 2 0.2537 0.904 0.000 0.920 0.080
#> SRR1274566 3 0.4235 0.804 0.000 0.176 0.824
#> SRR1274567 3 0.2878 0.885 0.000 0.096 0.904
#> SRR1274568 3 0.3038 0.886 0.000 0.104 0.896
#> SRR1274569 1 0.3267 0.925 0.884 0.000 0.116
#> SRR1274570 1 0.3192 0.925 0.888 0.000 0.112
#> SRR1274571 1 0.3267 0.925 0.884 0.000 0.116
#> SRR1274572 1 0.3192 0.925 0.888 0.000 0.112
#> SRR1274573 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274575 2 0.0592 0.966 0.000 0.988 0.012
#> SRR1274576 2 0.0000 0.972 0.000 1.000 0.000
#> SRR1274577 3 0.2356 0.890 0.000 0.072 0.928
#> SRR1274578 1 0.3267 0.925 0.884 0.000 0.116
#> SRR1274579 1 0.3267 0.925 0.884 0.000 0.116
#> SRR1274580 1 0.3267 0.925 0.884 0.000 0.116
#> SRR1274581 1 0.2066 0.913 0.940 0.000 0.060
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.4567 0.71436 0.000 0.740 0.244 0.016
#> SRR1274530 2 0.4661 0.70090 0.000 0.728 0.256 0.016
#> SRR1274531 4 0.3545 0.67268 0.000 0.008 0.164 0.828
#> SRR1274532 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.1109 0.81929 0.004 0.028 0.000 0.968
#> SRR1274534 4 0.1109 0.81929 0.004 0.028 0.000 0.968
#> SRR1274535 4 0.1151 0.80494 0.024 0.000 0.008 0.968
#> SRR1274536 4 0.4955 0.28815 0.000 0.008 0.344 0.648
#> SRR1274537 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.2647 0.82096 0.000 0.880 0.120 0.000
#> SRR1274539 4 0.1004 0.82170 0.004 0.024 0.000 0.972
#> SRR1274540 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.2589 0.82235 0.000 0.884 0.116 0.000
#> SRR1274544 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.3695 0.79442 0.000 0.828 0.156 0.016
#> SRR1274546 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.3024 0.80825 0.000 0.852 0.148 0.000
#> SRR1274548 2 0.5200 0.70654 0.000 0.744 0.184 0.072
#> SRR1274549 4 0.6469 0.44884 0.164 0.000 0.192 0.644
#> SRR1274550 1 0.4989 0.68384 0.528 0.000 0.472 0.000
#> SRR1274551 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274552 1 0.4993 0.56856 0.712 0.000 0.028 0.260
#> SRR1274553 1 0.5003 0.71849 0.768 0.000 0.148 0.084
#> SRR1274554 4 0.1109 0.81929 0.004 0.028 0.000 0.968
#> SRR1274555 1 0.4989 0.68384 0.528 0.000 0.472 0.000
#> SRR1274556 1 0.5000 0.66808 0.500 0.000 0.500 0.000
#> SRR1274557 1 0.4746 0.70723 0.632 0.000 0.368 0.000
#> SRR1274558 2 0.7301 -0.24999 0.000 0.452 0.152 0.396
#> SRR1274559 4 0.0524 0.81368 0.004 0.008 0.000 0.988
#> SRR1274560 4 0.2699 0.77668 0.028 0.000 0.068 0.904
#> SRR1274561 1 0.6137 -0.00967 0.504 0.000 0.048 0.448
#> SRR1274562 1 0.4985 0.68471 0.532 0.000 0.468 0.000
#> SRR1274563 1 0.4989 0.68384 0.528 0.000 0.472 0.000
#> SRR1274564 3 0.7529 0.64235 0.000 0.224 0.488 0.288
#> SRR1274565 2 0.6616 0.40970 0.000 0.584 0.308 0.108
#> SRR1274566 3 0.6495 0.50206 0.000 0.072 0.492 0.436
#> SRR1274567 4 0.4955 0.28815 0.000 0.008 0.344 0.648
#> SRR1274568 4 0.1004 0.82170 0.004 0.024 0.000 0.972
#> SRR1274569 1 0.0921 0.73614 0.972 0.000 0.000 0.028
#> SRR1274570 1 0.0817 0.73651 0.976 0.000 0.000 0.024
#> SRR1274571 1 0.0921 0.73614 0.972 0.000 0.000 0.028
#> SRR1274572 1 0.0817 0.73651 0.976 0.000 0.000 0.024
#> SRR1274573 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.84990 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.4661 0.70090 0.000 0.728 0.256 0.016
#> SRR1274576 2 0.3384 0.81352 0.000 0.860 0.116 0.024
#> SRR1274577 4 0.1004 0.82170 0.004 0.024 0.000 0.972
#> SRR1274578 1 0.0921 0.73614 0.972 0.000 0.000 0.028
#> SRR1274579 1 0.0921 0.73614 0.972 0.000 0.000 0.028
#> SRR1274580 1 0.0921 0.73614 0.972 0.000 0.000 0.028
#> SRR1274581 1 0.5296 0.66803 0.500 0.000 0.492 0.008
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0609 0.811 0.000 0.980 0.020 0.000 0.000
#> SRR1274529 2 0.6225 0.563 0.000 0.544 0.200 0.256 0.000
#> SRR1274530 2 0.6286 0.554 0.000 0.532 0.204 0.264 0.000
#> SRR1274531 5 0.3790 0.420 0.000 0.000 0.004 0.272 0.724
#> SRR1274532 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.0000 0.812 0.000 0.000 0.000 0.000 1.000
#> SRR1274534 5 0.0000 0.812 0.000 0.000 0.000 0.000 1.000
#> SRR1274535 5 0.0404 0.803 0.000 0.000 0.000 0.012 0.988
#> SRR1274536 4 0.3932 0.566 0.000 0.000 0.000 0.672 0.328
#> SRR1274537 2 0.0703 0.810 0.000 0.976 0.024 0.000 0.000
#> SRR1274538 2 0.4020 0.766 0.000 0.796 0.108 0.096 0.000
#> SRR1274539 5 0.0000 0.812 0.000 0.000 0.000 0.000 1.000
#> SRR1274540 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.3967 0.768 0.000 0.800 0.108 0.092 0.000
#> SRR1274544 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.4883 0.721 0.000 0.708 0.200 0.092 0.000
#> SRR1274546 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.5082 0.711 0.000 0.684 0.220 0.096 0.000
#> SRR1274548 2 0.7454 0.537 0.000 0.536 0.164 0.144 0.156
#> SRR1274549 4 0.6317 0.172 0.144 0.000 0.004 0.496 0.356
#> SRR1274550 3 0.3876 0.904 0.316 0.000 0.684 0.000 0.000
#> SRR1274551 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 1 0.7145 0.376 0.548 0.000 0.084 0.140 0.228
#> SRR1274553 1 0.7125 0.350 0.576 0.000 0.136 0.136 0.152
#> SRR1274554 5 0.0000 0.812 0.000 0.000 0.000 0.000 1.000
#> SRR1274555 3 0.3876 0.904 0.316 0.000 0.684 0.000 0.000
#> SRR1274556 3 0.3480 0.857 0.248 0.000 0.752 0.000 0.000
#> SRR1274557 3 0.5376 0.642 0.424 0.000 0.520 0.056 0.000
#> SRR1274558 5 0.7694 -0.123 0.000 0.316 0.116 0.128 0.440
#> SRR1274559 5 0.0579 0.803 0.000 0.000 0.008 0.008 0.984
#> SRR1274560 5 0.3835 0.457 0.000 0.000 0.008 0.260 0.732
#> SRR1274561 1 0.7040 -0.023 0.368 0.000 0.008 0.320 0.304
#> SRR1274562 3 0.3949 0.895 0.332 0.000 0.668 0.000 0.000
#> SRR1274563 3 0.3876 0.904 0.316 0.000 0.684 0.000 0.000
#> SRR1274564 4 0.6469 0.558 0.000 0.064 0.132 0.628 0.176
#> SRR1274565 2 0.7607 0.243 0.000 0.388 0.168 0.372 0.072
#> SRR1274566 4 0.6195 0.577 0.000 0.036 0.128 0.632 0.204
#> SRR1274567 4 0.3932 0.566 0.000 0.000 0.000 0.672 0.328
#> SRR1274568 5 0.0290 0.809 0.000 0.000 0.008 0.000 0.992
#> SRR1274569 1 0.1671 0.720 0.924 0.000 0.000 0.076 0.000
#> SRR1274570 1 0.0609 0.738 0.980 0.000 0.000 0.020 0.000
#> SRR1274571 1 0.1410 0.728 0.940 0.000 0.000 0.060 0.000
#> SRR1274572 1 0.0000 0.737 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.814 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.6286 0.554 0.000 0.532 0.204 0.264 0.000
#> SRR1274576 2 0.5787 0.709 0.000 0.704 0.112 0.092 0.092
#> SRR1274577 5 0.0000 0.812 0.000 0.000 0.000 0.000 1.000
#> SRR1274578 1 0.0404 0.738 0.988 0.000 0.000 0.012 0.000
#> SRR1274579 1 0.0404 0.738 0.988 0.000 0.000 0.012 0.000
#> SRR1274580 1 0.0404 0.738 0.988 0.000 0.000 0.012 0.000
#> SRR1274581 3 0.5010 0.812 0.248 0.000 0.676 0.076 0.000
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.1053 0.84093 0.000 0.964 0.004 0.012 0.000 0.020
#> SRR1274529 6 0.5100 0.64194 0.000 0.328 0.056 0.020 0.000 0.596
#> SRR1274530 6 0.4965 0.66210 0.000 0.296 0.036 0.036 0.000 0.632
#> SRR1274531 4 0.4532 0.49560 0.004 0.000 0.000 0.508 0.464 0.024
#> SRR1274532 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0146 0.87136 0.000 0.004 0.000 0.000 0.996 0.000
#> SRR1274534 5 0.0146 0.87136 0.000 0.004 0.000 0.000 0.996 0.000
#> SRR1274535 5 0.0508 0.85961 0.004 0.000 0.000 0.012 0.984 0.000
#> SRR1274536 4 0.5470 0.64831 0.000 0.000 0.016 0.620 0.160 0.204
#> SRR1274537 2 0.1218 0.83453 0.000 0.956 0.004 0.012 0.000 0.028
#> SRR1274538 2 0.4419 0.33237 0.000 0.656 0.028 0.012 0.000 0.304
#> SRR1274539 5 0.0146 0.87009 0.004 0.000 0.000 0.000 0.996 0.000
#> SRR1274540 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.4401 0.34365 0.000 0.660 0.028 0.012 0.000 0.300
#> SRR1274544 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 6 0.5238 0.49476 0.000 0.436 0.056 0.016 0.000 0.492
#> SRR1274546 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 6 0.5088 0.51939 0.000 0.412 0.036 0.024 0.000 0.528
#> SRR1274548 6 0.6709 0.27491 0.000 0.384 0.040 0.016 0.140 0.420
#> SRR1274549 4 0.4348 0.66501 0.064 0.000 0.000 0.748 0.164 0.024
#> SRR1274550 3 0.2527 0.90764 0.168 0.000 0.832 0.000 0.000 0.000
#> SRR1274551 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 1 0.7786 0.22520 0.380 0.000 0.072 0.312 0.176 0.060
#> SRR1274553 1 0.8017 0.24148 0.384 0.000 0.168 0.276 0.112 0.060
#> SRR1274554 5 0.0436 0.86946 0.000 0.004 0.000 0.004 0.988 0.004
#> SRR1274555 3 0.2527 0.90764 0.168 0.000 0.832 0.000 0.000 0.000
#> SRR1274556 3 0.2488 0.88067 0.124 0.000 0.864 0.008 0.004 0.000
#> SRR1274557 3 0.4632 0.70186 0.276 0.000 0.656 0.064 0.000 0.004
#> SRR1274558 5 0.6854 -0.04000 0.000 0.168 0.044 0.016 0.448 0.324
#> SRR1274559 5 0.1602 0.84405 0.004 0.000 0.016 0.020 0.944 0.016
#> SRR1274560 4 0.4239 0.49624 0.004 0.000 0.004 0.536 0.452 0.004
#> SRR1274561 4 0.5071 0.53657 0.172 0.000 0.004 0.664 0.156 0.004
#> SRR1274562 3 0.3012 0.89198 0.196 0.000 0.796 0.008 0.000 0.000
#> SRR1274563 3 0.2527 0.90764 0.168 0.000 0.832 0.000 0.000 0.000
#> SRR1274564 6 0.5228 0.27782 0.000 0.012 0.024 0.224 0.072 0.668
#> SRR1274565 6 0.5053 0.62306 0.000 0.196 0.048 0.044 0.012 0.700
#> SRR1274566 6 0.5036 0.23764 0.000 0.000 0.024 0.232 0.080 0.664
#> SRR1274567 4 0.5470 0.64831 0.000 0.000 0.016 0.620 0.160 0.204
#> SRR1274568 5 0.1261 0.85194 0.004 0.000 0.028 0.004 0.956 0.008
#> SRR1274569 1 0.2053 0.76947 0.888 0.000 0.000 0.108 0.000 0.004
#> SRR1274570 1 0.0937 0.79861 0.960 0.000 0.000 0.040 0.000 0.000
#> SRR1274571 1 0.1267 0.79435 0.940 0.000 0.000 0.060 0.000 0.000
#> SRR1274572 1 0.0000 0.79952 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.86378 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 6 0.4965 0.66210 0.000 0.296 0.036 0.036 0.000 0.632
#> SRR1274576 2 0.6389 -0.00487 0.000 0.516 0.040 0.012 0.124 0.308
#> SRR1274577 5 0.0291 0.87056 0.004 0.000 0.004 0.000 0.992 0.000
#> SRR1274578 1 0.0858 0.79942 0.968 0.000 0.000 0.004 0.000 0.028
#> SRR1274579 1 0.0858 0.79942 0.968 0.000 0.000 0.004 0.000 0.028
#> SRR1274580 1 0.0858 0.79942 0.968 0.000 0.000 0.004 0.000 0.028
#> SRR1274581 3 0.5092 0.77700 0.120 0.000 0.712 0.084 0.000 0.084
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "skmeans"]
# you can also extract it by
# res = res_list["ATC:skmeans"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'skmeans' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 4.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.957 0.984 0.5066 0.497 0.497
#> 3 3 0.976 0.951 0.975 0.2101 0.875 0.752
#> 4 4 0.929 0.903 0.959 0.1219 0.906 0.762
#> 5 5 0.832 0.828 0.860 0.0702 0.916 0.737
#> 6 6 0.817 0.776 0.861 0.0339 0.980 0.923
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 4
#> attr(,"optional")
#> [1] 2 3
There is also optional best \(k\) = 2 3 that is worth to check.
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.969 0.000 1.000
#> SRR1274529 2 0.000 0.969 0.000 1.000
#> SRR1274530 2 0.000 0.969 0.000 1.000
#> SRR1274531 1 0.000 1.000 1.000 0.000
#> SRR1274532 2 0.000 0.969 0.000 1.000
#> SRR1274533 2 0.327 0.913 0.060 0.940
#> SRR1274534 2 0.955 0.412 0.376 0.624
#> SRR1274535 1 0.000 1.000 1.000 0.000
#> SRR1274536 2 0.000 0.969 0.000 1.000
#> SRR1274537 2 0.000 0.969 0.000 1.000
#> SRR1274538 2 0.000 0.969 0.000 1.000
#> SRR1274539 1 0.000 1.000 1.000 0.000
#> SRR1274540 2 0.000 0.969 0.000 1.000
#> SRR1274541 2 0.000 0.969 0.000 1.000
#> SRR1274542 2 0.000 0.969 0.000 1.000
#> SRR1274543 2 0.000 0.969 0.000 1.000
#> SRR1274544 2 0.000 0.969 0.000 1.000
#> SRR1274545 2 0.000 0.969 0.000 1.000
#> SRR1274546 2 0.000 0.969 0.000 1.000
#> SRR1274547 2 0.000 0.969 0.000 1.000
#> SRR1274548 2 0.000 0.969 0.000 1.000
#> SRR1274549 1 0.000 1.000 1.000 0.000
#> SRR1274550 1 0.000 1.000 1.000 0.000
#> SRR1274551 2 0.000 0.969 0.000 1.000
#> SRR1274552 1 0.000 1.000 1.000 0.000
#> SRR1274553 1 0.000 1.000 1.000 0.000
#> SRR1274554 2 0.000 0.969 0.000 1.000
#> SRR1274555 1 0.000 1.000 1.000 0.000
#> SRR1274556 1 0.000 1.000 1.000 0.000
#> SRR1274557 1 0.000 1.000 1.000 0.000
#> SRR1274558 2 0.000 0.969 0.000 1.000
#> SRR1274559 1 0.000 1.000 1.000 0.000
#> SRR1274560 1 0.000 1.000 1.000 0.000
#> SRR1274561 1 0.000 1.000 1.000 0.000
#> SRR1274562 1 0.000 1.000 1.000 0.000
#> SRR1274563 1 0.000 1.000 1.000 0.000
#> SRR1274564 2 0.000 0.969 0.000 1.000
#> SRR1274565 2 0.000 0.969 0.000 1.000
#> SRR1274566 2 0.000 0.969 0.000 1.000
#> SRR1274567 2 0.000 0.969 0.000 1.000
#> SRR1274568 2 0.992 0.201 0.448 0.552
#> SRR1274569 1 0.000 1.000 1.000 0.000
#> SRR1274570 1 0.000 1.000 1.000 0.000
#> SRR1274571 1 0.000 1.000 1.000 0.000
#> SRR1274572 1 0.000 1.000 1.000 0.000
#> SRR1274573 2 0.000 0.969 0.000 1.000
#> SRR1274574 2 0.000 0.969 0.000 1.000
#> SRR1274575 2 0.000 0.969 0.000 1.000
#> SRR1274576 2 0.000 0.969 0.000 1.000
#> SRR1274577 1 0.000 1.000 1.000 0.000
#> SRR1274578 1 0.000 1.000 1.000 0.000
#> SRR1274579 1 0.000 1.000 1.000 0.000
#> SRR1274580 1 0.000 1.000 1.000 0.000
#> SRR1274581 1 0.000 1.000 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274531 1 0.6026 0.397 0.624 0.000 0.376
#> SRR1274532 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274533 3 0.1267 0.920 0.004 0.024 0.972
#> SRR1274534 3 0.0983 0.920 0.004 0.016 0.980
#> SRR1274535 3 0.4842 0.712 0.224 0.000 0.776
#> SRR1274536 2 0.1163 0.973 0.000 0.972 0.028
#> SRR1274537 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274538 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274539 3 0.0892 0.915 0.020 0.000 0.980
#> SRR1274540 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274548 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274549 1 0.1289 0.952 0.968 0.000 0.032
#> SRR1274550 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274551 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274552 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274553 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274554 3 0.1411 0.915 0.000 0.036 0.964
#> SRR1274555 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274556 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274557 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274558 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274559 1 0.4346 0.757 0.816 0.000 0.184
#> SRR1274560 1 0.1031 0.946 0.976 0.000 0.024
#> SRR1274561 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274562 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274563 1 0.0747 0.960 0.984 0.000 0.016
#> SRR1274564 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274565 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274566 2 0.0747 0.984 0.000 0.984 0.016
#> SRR1274567 2 0.1163 0.973 0.000 0.972 0.028
#> SRR1274568 3 0.4411 0.816 0.016 0.140 0.844
#> SRR1274569 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274576 2 0.0000 0.997 0.000 1.000 0.000
#> SRR1274577 3 0.1289 0.912 0.032 0.000 0.968
#> SRR1274578 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274579 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.961 1.000 0.000 0.000
#> SRR1274581 1 0.0747 0.960 0.984 0.000 0.016
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0188 0.996 0.000 0.996 0.000 0.004
#> SRR1274530 2 0.0188 0.996 0.000 0.996 0.000 0.004
#> SRR1274531 4 0.0804 0.718 0.008 0.000 0.012 0.980
#> SRR1274532 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.0000 0.877 0.000 0.000 1.000 0.000
#> SRR1274534 3 0.0000 0.877 0.000 0.000 1.000 0.000
#> SRR1274535 3 0.3873 0.633 0.228 0.000 0.772 0.000
#> SRR1274536 4 0.0000 0.724 0.000 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274539 3 0.0000 0.877 0.000 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274549 4 0.4866 0.289 0.404 0.000 0.000 0.596
#> SRR1274550 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274551 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274552 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274553 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274554 3 0.0336 0.871 0.000 0.008 0.992 0.000
#> SRR1274555 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274556 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274557 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274558 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274559 1 0.3688 0.734 0.792 0.000 0.208 0.000
#> SRR1274560 1 0.3873 0.685 0.772 0.000 0.000 0.228
#> SRR1274561 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274562 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274563 1 0.1211 0.950 0.960 0.000 0.040 0.000
#> SRR1274564 4 0.4454 0.499 0.000 0.308 0.000 0.692
#> SRR1274565 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274566 4 0.2647 0.683 0.000 0.120 0.000 0.880
#> SRR1274567 4 0.0000 0.724 0.000 0.000 0.000 1.000
#> SRR1274568 3 0.5610 0.583 0.032 0.008 0.668 0.292
#> SRR1274569 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.0188 0.996 0.000 0.996 0.000 0.004
#> SRR1274576 2 0.0000 0.999 0.000 1.000 0.000 0.000
#> SRR1274577 3 0.1211 0.857 0.040 0.000 0.960 0.000
#> SRR1274578 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.953 1.000 0.000 0.000 0.000
#> SRR1274581 1 0.1211 0.950 0.960 0.000 0.040 0.000
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.2771 0.866 0.012 0.860 0.000 0.128 0.000
#> SRR1274530 2 0.2909 0.854 0.012 0.848 0.000 0.140 0.000
#> SRR1274531 4 0.6739 0.404 0.224 0.000 0.288 0.480 0.008
#> SRR1274532 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.0510 0.872 0.016 0.000 0.000 0.000 0.984
#> SRR1274534 5 0.0162 0.874 0.004 0.000 0.000 0.000 0.996
#> SRR1274535 3 0.5142 0.463 0.088 0.000 0.668 0.000 0.244
#> SRR1274536 4 0.1121 0.798 0.044 0.000 0.000 0.956 0.000
#> SRR1274537 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.0510 0.873 0.016 0.000 0.000 0.000 0.984
#> SRR1274540 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.1522 0.938 0.012 0.944 0.000 0.044 0.000
#> SRR1274546 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.0324 0.968 0.004 0.992 0.000 0.004 0.000
#> SRR1274548 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274549 3 0.5174 0.335 0.056 0.000 0.604 0.340 0.000
#> SRR1274550 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274551 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274553 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274554 5 0.0703 0.869 0.024 0.000 0.000 0.000 0.976
#> SRR1274555 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274559 1 0.6396 0.444 0.512 0.000 0.264 0.000 0.224
#> SRR1274560 1 0.3810 0.583 0.792 0.000 0.168 0.040 0.000
#> SRR1274561 1 0.4300 0.763 0.524 0.000 0.476 0.000 0.000
#> SRR1274562 3 0.3586 0.152 0.264 0.000 0.736 0.000 0.000
#> SRR1274563 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.2416 0.702 0.012 0.100 0.000 0.888 0.000
#> SRR1274565 2 0.1877 0.924 0.012 0.924 0.000 0.064 0.000
#> SRR1274566 4 0.0693 0.786 0.012 0.008 0.000 0.980 0.000
#> SRR1274567 4 0.1121 0.798 0.044 0.000 0.000 0.956 0.000
#> SRR1274568 5 0.5605 0.537 0.404 0.000 0.000 0.076 0.520
#> SRR1274569 1 0.4278 0.799 0.548 0.000 0.452 0.000 0.000
#> SRR1274570 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274571 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274572 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274573 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 2 0.2909 0.854 0.012 0.848 0.000 0.140 0.000
#> SRR1274576 2 0.0000 0.973 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 5 0.3427 0.788 0.192 0.000 0.012 0.000 0.796
#> SRR1274578 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274579 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274580 1 0.4182 0.863 0.600 0.000 0.400 0.000 0.000
#> SRR1274581 3 0.0000 0.839 0.000 0.000 1.000 0.000 0.000
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.4107 0.7778 0.000 0.768 0.024 0.052 0.000 0.156
#> SRR1274530 2 0.4202 0.7617 0.000 0.756 0.012 0.080 0.000 0.152
#> SRR1274531 4 0.5947 0.0802 0.000 0.000 0.404 0.408 0.004 0.184
#> SRR1274532 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.1003 0.8775 0.000 0.000 0.016 0.000 0.964 0.020
#> SRR1274534 5 0.1408 0.8875 0.000 0.000 0.020 0.000 0.944 0.036
#> SRR1274535 3 0.5668 0.5476 0.088 0.000 0.656 0.000 0.140 0.116
#> SRR1274536 4 0.0146 0.6797 0.000 0.000 0.000 0.996 0.000 0.004
#> SRR1274537 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.1845 0.8727 0.000 0.000 0.028 0.000 0.920 0.052
#> SRR1274540 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.2730 0.8414 0.000 0.836 0.012 0.000 0.000 0.152
#> SRR1274546 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.1007 0.9243 0.000 0.956 0.000 0.000 0.000 0.044
#> SRR1274548 2 0.0458 0.9392 0.000 0.984 0.000 0.000 0.000 0.016
#> SRR1274549 3 0.5876 0.3720 0.160 0.000 0.476 0.356 0.000 0.008
#> SRR1274550 3 0.2854 0.9022 0.208 0.000 0.792 0.000 0.000 0.000
#> SRR1274551 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 3 0.2994 0.9008 0.208 0.000 0.788 0.000 0.000 0.004
#> SRR1274553 3 0.2994 0.9008 0.208 0.000 0.788 0.000 0.000 0.004
#> SRR1274554 5 0.1713 0.8539 0.000 0.000 0.028 0.000 0.928 0.044
#> SRR1274555 3 0.2854 0.9022 0.208 0.000 0.792 0.000 0.000 0.000
#> SRR1274556 3 0.2854 0.9022 0.208 0.000 0.792 0.000 0.000 0.000
#> SRR1274557 3 0.2854 0.9022 0.208 0.000 0.792 0.000 0.000 0.000
#> SRR1274558 2 0.0260 0.9423 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274559 1 0.7388 0.0523 0.376 0.000 0.228 0.000 0.264 0.132
#> SRR1274560 1 0.4901 0.5124 0.708 0.000 0.100 0.032 0.000 0.160
#> SRR1274561 1 0.2814 0.6513 0.820 0.000 0.172 0.000 0.000 0.008
#> SRR1274562 1 0.3868 -0.3356 0.504 0.000 0.496 0.000 0.000 0.000
#> SRR1274563 3 0.2854 0.9022 0.208 0.000 0.792 0.000 0.000 0.000
#> SRR1274564 4 0.3881 0.6247 0.000 0.040 0.024 0.784 0.000 0.152
#> SRR1274565 2 0.3399 0.8245 0.000 0.816 0.024 0.020 0.000 0.140
#> SRR1274566 4 0.3084 0.6612 0.000 0.008 0.024 0.832 0.000 0.136
#> SRR1274567 4 0.0146 0.6820 0.000 0.000 0.000 0.996 0.000 0.004
#> SRR1274568 6 0.4376 0.6277 0.036 0.000 0.004 0.056 0.140 0.764
#> SRR1274569 1 0.1967 0.7497 0.904 0.000 0.084 0.000 0.000 0.012
#> SRR1274570 1 0.0622 0.7987 0.980 0.000 0.008 0.000 0.000 0.012
#> SRR1274571 1 0.0622 0.7987 0.980 0.000 0.008 0.000 0.000 0.012
#> SRR1274572 1 0.0260 0.7999 0.992 0.000 0.008 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.9462 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 2 0.4202 0.7617 0.000 0.756 0.012 0.080 0.000 0.152
#> SRR1274576 2 0.0260 0.9424 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274577 6 0.4426 0.5039 0.004 0.000 0.024 0.000 0.388 0.584
#> SRR1274578 1 0.0458 0.7996 0.984 0.000 0.000 0.000 0.000 0.016
#> SRR1274579 1 0.0458 0.7996 0.984 0.000 0.000 0.000 0.000 0.016
#> SRR1274580 1 0.0458 0.7996 0.984 0.000 0.000 0.000 0.000 0.016
#> SRR1274581 3 0.2883 0.8991 0.212 0.000 0.788 0.000 0.000 0.000
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "pam"]
# you can also extract it by
# res = res_list["ATC:pam"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'pam' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 6.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.823 0.888 0.957 0.4803 0.525 0.525
#> 3 3 0.929 0.913 0.966 0.4073 0.739 0.530
#> 4 4 0.985 0.921 0.968 0.0842 0.916 0.750
#> 5 5 0.857 0.792 0.919 0.0692 0.915 0.697
#> 6 6 0.908 0.829 0.937 0.0202 0.963 0.831
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 6
#> attr(,"optional")
#> [1] 3 4
There is also optional best \(k\) = 3 4 that is worth to check.
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.944 0.000 1.000
#> SRR1274529 2 0.000 0.944 0.000 1.000
#> SRR1274530 2 0.000 0.944 0.000 1.000
#> SRR1274531 2 0.969 0.375 0.396 0.604
#> SRR1274532 2 0.000 0.944 0.000 1.000
#> SRR1274533 2 0.118 0.935 0.016 0.984
#> SRR1274534 2 0.118 0.935 0.016 0.984
#> SRR1274535 1 0.983 0.175 0.576 0.424
#> SRR1274536 2 0.969 0.375 0.396 0.604
#> SRR1274537 2 0.000 0.944 0.000 1.000
#> SRR1274538 2 0.000 0.944 0.000 1.000
#> SRR1274539 2 0.184 0.927 0.028 0.972
#> SRR1274540 2 0.000 0.944 0.000 1.000
#> SRR1274541 2 0.000 0.944 0.000 1.000
#> SRR1274542 2 0.000 0.944 0.000 1.000
#> SRR1274543 2 0.000 0.944 0.000 1.000
#> SRR1274544 2 0.000 0.944 0.000 1.000
#> SRR1274545 2 0.000 0.944 0.000 1.000
#> SRR1274546 2 0.000 0.944 0.000 1.000
#> SRR1274547 2 0.000 0.944 0.000 1.000
#> SRR1274548 2 0.000 0.944 0.000 1.000
#> SRR1274549 1 0.000 0.964 1.000 0.000
#> SRR1274550 1 0.000 0.964 1.000 0.000
#> SRR1274551 2 0.000 0.944 0.000 1.000
#> SRR1274552 1 0.000 0.964 1.000 0.000
#> SRR1274553 1 0.000 0.964 1.000 0.000
#> SRR1274554 2 0.118 0.935 0.016 0.984
#> SRR1274555 1 0.000 0.964 1.000 0.000
#> SRR1274556 1 0.000 0.964 1.000 0.000
#> SRR1274557 1 0.000 0.964 1.000 0.000
#> SRR1274558 2 0.000 0.944 0.000 1.000
#> SRR1274559 2 0.969 0.375 0.396 0.604
#> SRR1274560 1 0.722 0.717 0.800 0.200
#> SRR1274561 1 0.000 0.964 1.000 0.000
#> SRR1274562 1 0.000 0.964 1.000 0.000
#> SRR1274563 1 0.000 0.964 1.000 0.000
#> SRR1274564 2 0.000 0.944 0.000 1.000
#> SRR1274565 2 0.000 0.944 0.000 1.000
#> SRR1274566 2 0.000 0.944 0.000 1.000
#> SRR1274567 2 0.969 0.375 0.396 0.604
#> SRR1274568 2 0.184 0.927 0.028 0.972
#> SRR1274569 1 0.000 0.964 1.000 0.000
#> SRR1274570 1 0.000 0.964 1.000 0.000
#> SRR1274571 1 0.000 0.964 1.000 0.000
#> SRR1274572 1 0.000 0.964 1.000 0.000
#> SRR1274573 2 0.000 0.944 0.000 1.000
#> SRR1274574 2 0.000 0.944 0.000 1.000
#> SRR1274575 2 0.000 0.944 0.000 1.000
#> SRR1274576 2 0.000 0.944 0.000 1.000
#> SRR1274577 2 0.163 0.930 0.024 0.976
#> SRR1274578 1 0.000 0.964 1.000 0.000
#> SRR1274579 1 0.000 0.964 1.000 0.000
#> SRR1274580 1 0.000 0.964 1.000 0.000
#> SRR1274581 1 0.000 0.964 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274529 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274530 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274531 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274533 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274534 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274535 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274536 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274538 2 0.0237 0.996 0.000 0.996 0.004
#> SRR1274539 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274540 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274543 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274544 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274545 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274546 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274547 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274548 2 0.0592 0.988 0.000 0.988 0.012
#> SRR1274549 3 0.6244 0.101 0.440 0.000 0.560
#> SRR1274550 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274551 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274552 1 0.6126 0.382 0.600 0.000 0.400
#> SRR1274553 1 0.0892 0.920 0.980 0.000 0.020
#> SRR1274554 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274555 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274556 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274557 1 0.6095 0.402 0.608 0.000 0.392
#> SRR1274558 3 0.0592 0.933 0.000 0.012 0.988
#> SRR1274559 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274560 3 0.4796 0.692 0.220 0.000 0.780
#> SRR1274561 1 0.0424 0.926 0.992 0.000 0.008
#> SRR1274562 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274563 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274564 3 0.3619 0.804 0.000 0.136 0.864
#> SRR1274565 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274566 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274567 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274568 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274569 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274570 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274571 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274572 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274573 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274575 2 0.0000 0.999 0.000 1.000 0.000
#> SRR1274576 2 0.0237 0.996 0.000 0.996 0.004
#> SRR1274577 3 0.0000 0.943 0.000 0.000 1.000
#> SRR1274578 1 0.0747 0.922 0.984 0.000 0.016
#> SRR1274579 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274580 1 0.0000 0.931 1.000 0.000 0.000
#> SRR1274581 1 0.4504 0.743 0.804 0.000 0.196
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274529 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274530 2 0.2081 0.905 0.000 0.916 0.000 0.084
#> SRR1274531 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274532 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274533 4 0.0188 0.974 0.000 0.004 0.000 0.996
#> SRR1274534 4 0.0188 0.974 0.000 0.004 0.000 0.996
#> SRR1274535 4 0.2081 0.898 0.000 0.000 0.084 0.916
#> SRR1274536 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274537 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274538 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274539 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274540 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274543 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274544 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274546 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274547 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274548 2 0.0336 0.982 0.000 0.992 0.000 0.008
#> SRR1274549 1 0.2412 0.839 0.908 0.000 0.008 0.084
#> SRR1274550 3 0.0000 0.911 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274552 1 0.6538 0.500 0.600 0.000 0.108 0.292
#> SRR1274553 3 0.4955 0.142 0.444 0.000 0.556 0.000
#> SRR1274554 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274555 3 0.0000 0.911 0.000 0.000 1.000 0.000
#> SRR1274556 3 0.0000 0.911 0.000 0.000 1.000 0.000
#> SRR1274557 3 0.0592 0.903 0.016 0.000 0.984 0.000
#> SRR1274558 4 0.2281 0.877 0.000 0.096 0.000 0.904
#> SRR1274559 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274560 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274561 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274562 3 0.2469 0.825 0.108 0.000 0.892 0.000
#> SRR1274563 3 0.0000 0.911 0.000 0.000 1.000 0.000
#> SRR1274564 4 0.2081 0.885 0.000 0.084 0.000 0.916
#> SRR1274565 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274566 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274567 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274568 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274569 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.2081 0.905 0.000 0.916 0.000 0.084
#> SRR1274576 2 0.0000 0.989 0.000 1.000 0.000 0.000
#> SRR1274577 4 0.0000 0.976 0.000 0.000 0.000 1.000
#> SRR1274578 1 0.3873 0.693 0.772 0.000 0.000 0.228
#> SRR1274579 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.912 1.000 0.000 0.000 0.000
#> SRR1274581 3 0.0000 0.911 0.000 0.000 1.000 0.000
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274529 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274530 4 0.429 0.2276 0.000 0.468 0.000 0.532 0.000
#> SRR1274531 4 0.273 0.6744 0.000 0.000 0.000 0.840 0.160
#> SRR1274532 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274533 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274534 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274535 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274536 4 0.000 0.8239 0.000 0.000 0.000 1.000 0.000
#> SRR1274537 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274538 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274539 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274540 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274541 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274542 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274543 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274544 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274545 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274546 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274547 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274548 2 0.348 0.6416 0.000 0.752 0.000 0.248 0.000
#> SRR1274549 1 0.632 0.1413 0.444 0.000 0.000 0.400 0.156
#> SRR1274550 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
#> SRR1274551 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274552 5 0.590 -0.1993 0.444 0.000 0.100 0.000 0.456
#> SRR1274553 3 0.427 0.2024 0.444 0.000 0.556 0.000 0.000
#> SRR1274554 5 0.269 0.6620 0.000 0.000 0.000 0.156 0.844
#> SRR1274555 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
#> SRR1274556 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
#> SRR1274557 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
#> SRR1274558 5 0.637 0.1859 0.000 0.400 0.000 0.164 0.436
#> SRR1274559 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274560 1 0.269 0.7805 0.844 0.000 0.000 0.000 0.156
#> SRR1274561 1 0.269 0.7805 0.844 0.000 0.000 0.000 0.156
#> SRR1274562 3 0.202 0.8323 0.100 0.000 0.900 0.000 0.000
#> SRR1274563 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
#> SRR1274564 4 0.000 0.8239 0.000 0.000 0.000 1.000 0.000
#> SRR1274565 4 0.000 0.8239 0.000 0.000 0.000 1.000 0.000
#> SRR1274566 4 0.000 0.8239 0.000 0.000 0.000 1.000 0.000
#> SRR1274567 4 0.000 0.8239 0.000 0.000 0.000 1.000 0.000
#> SRR1274568 5 0.430 0.0695 0.000 0.000 0.000 0.484 0.516
#> SRR1274569 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274574 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274575 4 0.337 0.6434 0.000 0.232 0.000 0.768 0.000
#> SRR1274576 2 0.000 0.9836 0.000 1.000 0.000 0.000 0.000
#> SRR1274577 5 0.000 0.7818 0.000 0.000 0.000 0.000 1.000
#> SRR1274578 1 0.191 0.8188 0.908 0.000 0.000 0.000 0.092
#> SRR1274579 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.000 0.8833 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.000 0.9134 0.000 0.000 1.000 0.000 0.000
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274529 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274530 4 0.3857 0.202 0.000 0.468 0.000 0.532 0.000 0.000
#> SRR1274531 4 0.0458 0.788 0.000 0.000 0.000 0.984 0.016 0.000
#> SRR1274532 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274533 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274534 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274535 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274536 4 0.0000 0.797 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274537 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274538 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274539 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274540 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274541 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274542 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274543 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274544 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274545 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274546 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274547 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274548 2 0.3126 0.638 0.000 0.752 0.000 0.248 0.000 0.000
#> SRR1274549 1 0.3993 0.333 0.592 0.000 0.000 0.400 0.008 0.000
#> SRR1274550 3 0.0000 0.900 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274551 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274552 1 0.5042 0.473 0.592 0.000 0.100 0.000 0.308 0.000
#> SRR1274553 3 0.3833 0.143 0.444 0.000 0.556 0.000 0.000 0.000
#> SRR1274554 5 0.0260 0.881 0.000 0.000 0.000 0.008 0.992 0.000
#> SRR1274555 3 0.0000 0.900 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.0000 0.900 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274557 3 0.0547 0.890 0.020 0.000 0.980 0.000 0.000 0.000
#> SRR1274558 5 0.5659 0.103 0.000 0.424 0.000 0.152 0.424 0.000
#> SRR1274559 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274560 1 0.0260 0.860 0.992 0.000 0.000 0.000 0.008 0.000
#> SRR1274561 1 0.0260 0.860 0.992 0.000 0.000 0.000 0.008 0.000
#> SRR1274562 3 0.1814 0.817 0.100 0.000 0.900 0.000 0.000 0.000
#> SRR1274563 3 0.0000 0.900 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 4 0.0000 0.797 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274565 4 0.0000 0.797 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274566 4 0.0000 0.797 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274567 4 0.0000 0.797 0.000 0.000 0.000 1.000 0.000 0.000
#> SRR1274568 4 0.3804 0.189 0.000 0.000 0.000 0.576 0.424 0.000
#> SRR1274569 1 0.0000 0.862 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274570 1 0.0000 0.862 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274571 1 0.0000 0.862 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274572 1 0.0000 0.862 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274573 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274574 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274575 4 0.3023 0.584 0.000 0.232 0.000 0.768 0.000 0.000
#> SRR1274576 2 0.0000 0.984 0.000 1.000 0.000 0.000 0.000 0.000
#> SRR1274577 5 0.0000 0.888 0.000 0.000 0.000 0.000 1.000 0.000
#> SRR1274578 6 0.0146 1.000 0.004 0.000 0.000 0.000 0.000 0.996
#> SRR1274579 6 0.0146 1.000 0.004 0.000 0.000 0.000 0.000 0.996
#> SRR1274580 6 0.0146 1.000 0.004 0.000 0.000 0.000 0.000 0.996
#> SRR1274581 3 0.0146 0.899 0.000 0.000 0.996 0.000 0.000 0.004
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "mclust"]
# you can also extract it by
# res = res_list["ATC:mclust"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'mclust' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 0.752 0.916 0.962 0.4702 0.535 0.535
#> 3 3 0.517 0.712 0.858 0.2427 0.755 0.603
#> 4 4 0.533 0.492 0.739 0.1802 0.813 0.597
#> 5 5 0.539 0.474 0.702 0.0914 0.742 0.380
#> 6 6 0.778 0.806 0.837 0.0753 0.840 0.475
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.0000 0.961 0.000 1.000
#> SRR1274529 2 0.0000 0.961 0.000 1.000
#> SRR1274530 2 0.7056 0.765 0.192 0.808
#> SRR1274531 1 0.0000 0.955 1.000 0.000
#> SRR1274532 2 0.0000 0.961 0.000 1.000
#> SRR1274533 1 0.0000 0.955 1.000 0.000
#> SRR1274534 1 0.0000 0.955 1.000 0.000
#> SRR1274535 1 0.0000 0.955 1.000 0.000
#> SRR1274536 1 0.8016 0.687 0.756 0.244
#> SRR1274537 2 0.0672 0.957 0.008 0.992
#> SRR1274538 2 0.0672 0.957 0.008 0.992
#> SRR1274539 1 0.0000 0.955 1.000 0.000
#> SRR1274540 2 0.0000 0.961 0.000 1.000
#> SRR1274541 2 0.0000 0.961 0.000 1.000
#> SRR1274542 2 0.0000 0.961 0.000 1.000
#> SRR1274543 2 0.0672 0.957 0.008 0.992
#> SRR1274544 2 0.0000 0.961 0.000 1.000
#> SRR1274545 2 0.0000 0.961 0.000 1.000
#> SRR1274546 2 0.0000 0.961 0.000 1.000
#> SRR1274547 2 0.0000 0.961 0.000 1.000
#> SRR1274548 1 0.0672 0.950 0.992 0.008
#> SRR1274549 1 0.0000 0.955 1.000 0.000
#> SRR1274550 1 0.0000 0.955 1.000 0.000
#> SRR1274551 2 0.0000 0.961 0.000 1.000
#> SRR1274552 1 0.0000 0.955 1.000 0.000
#> SRR1274553 1 0.0000 0.955 1.000 0.000
#> SRR1274554 1 0.0000 0.955 1.000 0.000
#> SRR1274555 1 0.0000 0.955 1.000 0.000
#> SRR1274556 1 0.0000 0.955 1.000 0.000
#> SRR1274557 1 0.0000 0.955 1.000 0.000
#> SRR1274558 1 0.6247 0.802 0.844 0.156
#> SRR1274559 1 0.0000 0.955 1.000 0.000
#> SRR1274560 1 0.0000 0.955 1.000 0.000
#> SRR1274561 1 0.0000 0.955 1.000 0.000
#> SRR1274562 1 0.0000 0.955 1.000 0.000
#> SRR1274563 1 0.0000 0.955 1.000 0.000
#> SRR1274564 1 0.8016 0.687 0.756 0.244
#> SRR1274565 2 0.8016 0.669 0.244 0.756
#> SRR1274566 1 0.8016 0.687 0.756 0.244
#> SRR1274567 1 0.8016 0.687 0.756 0.244
#> SRR1274568 1 0.0000 0.955 1.000 0.000
#> SRR1274569 1 0.0000 0.955 1.000 0.000
#> SRR1274570 1 0.0000 0.955 1.000 0.000
#> SRR1274571 1 0.0000 0.955 1.000 0.000
#> SRR1274572 1 0.0000 0.955 1.000 0.000
#> SRR1274573 2 0.0000 0.961 0.000 1.000
#> SRR1274574 2 0.0000 0.961 0.000 1.000
#> SRR1274575 2 0.7056 0.765 0.192 0.808
#> SRR1274576 1 0.8081 0.671 0.752 0.248
#> SRR1274577 1 0.0000 0.955 1.000 0.000
#> SRR1274578 1 0.0000 0.955 1.000 0.000
#> SRR1274579 1 0.0000 0.955 1.000 0.000
#> SRR1274580 1 0.0000 0.955 1.000 0.000
#> SRR1274581 1 0.0000 0.955 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.2625 0.8869 0.000 0.916 0.084
#> SRR1274529 2 0.2625 0.8868 0.000 0.916 0.084
#> SRR1274530 2 0.5580 0.7220 0.008 0.736 0.256
#> SRR1274531 3 0.3769 0.7594 0.104 0.016 0.880
#> SRR1274532 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274533 3 0.1015 0.7756 0.008 0.012 0.980
#> SRR1274534 3 0.1015 0.7756 0.008 0.012 0.980
#> SRR1274535 3 0.2939 0.7613 0.072 0.012 0.916
#> SRR1274536 3 0.4662 0.7424 0.124 0.032 0.844
#> SRR1274537 3 0.6295 0.0783 0.000 0.472 0.528
#> SRR1274538 3 0.5254 0.5620 0.000 0.264 0.736
#> SRR1274539 3 0.1751 0.7741 0.028 0.012 0.960
#> SRR1274540 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274542 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274543 3 0.6225 0.1534 0.000 0.432 0.568
#> SRR1274544 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274545 2 0.2625 0.8868 0.000 0.916 0.084
#> SRR1274546 2 0.0424 0.9116 0.000 0.992 0.008
#> SRR1274547 2 0.4887 0.7564 0.000 0.772 0.228
#> SRR1274548 3 0.1765 0.7694 0.004 0.040 0.956
#> SRR1274549 3 0.5070 0.7064 0.224 0.004 0.772
#> SRR1274550 3 0.5926 0.4042 0.356 0.000 0.644
#> SRR1274551 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274552 3 0.3618 0.7477 0.104 0.012 0.884
#> SRR1274553 3 0.5493 0.6246 0.232 0.012 0.756
#> SRR1274554 3 0.0592 0.7750 0.000 0.012 0.988
#> SRR1274555 3 0.5926 0.4042 0.356 0.000 0.644
#> SRR1274556 3 0.5926 0.4042 0.356 0.000 0.644
#> SRR1274557 3 0.3918 0.7389 0.120 0.012 0.868
#> SRR1274558 3 0.1031 0.7731 0.000 0.024 0.976
#> SRR1274559 3 0.0592 0.7750 0.000 0.012 0.988
#> SRR1274560 3 0.3879 0.7365 0.152 0.000 0.848
#> SRR1274561 3 0.4555 0.7075 0.200 0.000 0.800
#> SRR1274562 3 0.6168 0.3216 0.412 0.000 0.588
#> SRR1274563 3 0.5926 0.4042 0.356 0.000 0.644
#> SRR1274564 3 0.4121 0.7357 0.040 0.084 0.876
#> SRR1274565 3 0.3412 0.7225 0.000 0.124 0.876
#> SRR1274566 3 0.4092 0.7353 0.036 0.088 0.876
#> SRR1274567 3 0.4519 0.7424 0.116 0.032 0.852
#> SRR1274568 3 0.0592 0.7750 0.000 0.012 0.988
#> SRR1274569 3 0.5201 0.6759 0.236 0.004 0.760
#> SRR1274570 1 0.5621 0.5256 0.692 0.000 0.308
#> SRR1274571 1 0.5785 0.4830 0.668 0.000 0.332
#> SRR1274572 1 0.1163 0.8375 0.972 0.000 0.028
#> SRR1274573 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274574 2 0.0000 0.9132 0.000 1.000 0.000
#> SRR1274575 2 0.5580 0.7220 0.008 0.736 0.256
#> SRR1274576 3 0.1964 0.7607 0.000 0.056 0.944
#> SRR1274577 3 0.0829 0.7754 0.004 0.012 0.984
#> SRR1274578 1 0.1163 0.8375 0.972 0.000 0.028
#> SRR1274579 1 0.1163 0.8375 0.972 0.000 0.028
#> SRR1274580 1 0.1163 0.8375 0.972 0.000 0.028
#> SRR1274581 3 0.5926 0.4042 0.356 0.000 0.644
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.4039 0.7621 0.080 0.836 0.000 0.084
#> SRR1274529 2 0.5077 0.7380 0.080 0.760 0.000 0.160
#> SRR1274530 2 0.6546 0.4389 0.080 0.524 0.000 0.396
#> SRR1274531 4 0.5898 -0.3590 0.012 0.016 0.448 0.524
#> SRR1274532 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274533 3 0.5506 0.4901 0.000 0.016 0.512 0.472
#> SRR1274534 3 0.5500 0.4902 0.000 0.016 0.520 0.464
#> SRR1274535 3 0.5427 0.4640 0.000 0.016 0.568 0.416
#> SRR1274536 4 0.2408 0.5100 0.000 0.000 0.104 0.896
#> SRR1274537 2 0.8169 -0.2118 0.080 0.432 0.408 0.080
#> SRR1274538 3 0.9136 0.3071 0.080 0.244 0.404 0.272
#> SRR1274539 3 0.5427 0.4640 0.000 0.016 0.568 0.416
#> SRR1274540 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274541 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274542 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274543 3 0.8795 0.2197 0.080 0.372 0.400 0.148
#> SRR1274544 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274545 2 0.5032 0.7400 0.080 0.764 0.000 0.156
#> SRR1274546 2 0.0336 0.8112 0.000 0.992 0.000 0.008
#> SRR1274547 2 0.6201 0.5998 0.080 0.620 0.000 0.300
#> SRR1274548 3 0.8301 0.4069 0.076 0.096 0.424 0.404
#> SRR1274549 4 0.4212 0.5195 0.012 0.000 0.216 0.772
#> SRR1274550 3 0.0804 0.3658 0.012 0.000 0.980 0.008
#> SRR1274551 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274552 3 0.3893 0.3843 0.008 0.000 0.796 0.196
#> SRR1274553 3 0.3790 0.3767 0.016 0.000 0.820 0.164
#> SRR1274554 3 0.5506 0.4901 0.000 0.016 0.512 0.472
#> SRR1274555 3 0.0937 0.3651 0.012 0.000 0.976 0.012
#> SRR1274556 3 0.0469 0.3595 0.012 0.000 0.988 0.000
#> SRR1274557 3 0.2542 0.3800 0.012 0.000 0.904 0.084
#> SRR1274558 3 0.7118 0.4749 0.052 0.036 0.492 0.420
#> SRR1274559 3 0.5444 0.4955 0.000 0.016 0.560 0.424
#> SRR1274560 4 0.3852 0.4814 0.008 0.000 0.192 0.800
#> SRR1274561 4 0.5268 0.4128 0.012 0.000 0.396 0.592
#> SRR1274562 4 0.6705 0.1206 0.088 0.000 0.440 0.472
#> SRR1274563 3 0.0937 0.3651 0.012 0.000 0.976 0.012
#> SRR1274564 4 0.6703 -0.0480 0.036 0.052 0.292 0.620
#> SRR1274565 3 0.8944 0.3421 0.076 0.184 0.404 0.336
#> SRR1274566 4 0.4456 0.0304 0.000 0.004 0.280 0.716
#> SRR1274567 4 0.2408 0.5100 0.000 0.000 0.104 0.896
#> SRR1274568 3 0.5503 0.4866 0.000 0.016 0.516 0.468
#> SRR1274569 4 0.6552 0.2963 0.096 0.000 0.328 0.576
#> SRR1274570 1 0.7557 0.5008 0.484 0.000 0.284 0.232
#> SRR1274571 1 0.7593 0.4874 0.476 0.000 0.288 0.236
#> SRR1274572 1 0.2011 0.8163 0.920 0.000 0.080 0.000
#> SRR1274573 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274574 2 0.0000 0.8128 0.000 1.000 0.000 0.000
#> SRR1274575 2 0.6546 0.4389 0.080 0.524 0.000 0.396
#> SRR1274576 3 0.7374 0.4677 0.076 0.032 0.488 0.404
#> SRR1274577 3 0.5503 0.4906 0.000 0.016 0.516 0.468
#> SRR1274578 1 0.2011 0.8163 0.920 0.000 0.080 0.000
#> SRR1274579 1 0.2011 0.8163 0.920 0.000 0.080 0.000
#> SRR1274580 1 0.2011 0.8163 0.920 0.000 0.080 0.000
#> SRR1274581 3 0.0592 0.3571 0.016 0.000 0.984 0.000
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.2446 0.6066 0.000 0.900 0.000 0.044 0.056
#> SRR1274529 2 0.1043 0.6096 0.000 0.960 0.000 0.040 0.000
#> SRR1274530 2 0.2139 0.5809 0.000 0.916 0.000 0.032 0.052
#> SRR1274531 5 0.4536 0.1791 0.012 0.024 0.008 0.212 0.744
#> SRR1274532 2 0.4305 0.5713 0.000 0.512 0.000 0.488 0.000
#> SRR1274533 5 0.2248 0.5821 0.000 0.088 0.000 0.012 0.900
#> SRR1274534 5 0.2331 0.5792 0.000 0.080 0.000 0.020 0.900
#> SRR1274535 5 0.1788 0.5724 0.000 0.056 0.004 0.008 0.932
#> SRR1274536 5 0.6691 0.1341 0.004 0.180 0.016 0.256 0.544
#> SRR1274537 2 0.3988 0.5230 0.000 0.768 0.000 0.036 0.196
#> SRR1274538 2 0.3954 0.5153 0.000 0.772 0.000 0.036 0.192
#> SRR1274539 5 0.2074 0.5750 0.000 0.060 0.004 0.016 0.920
#> SRR1274540 2 0.4305 0.5713 0.000 0.512 0.000 0.488 0.000
#> SRR1274541 2 0.4450 0.5702 0.000 0.508 0.000 0.488 0.004
#> SRR1274542 2 0.4305 0.5713 0.000 0.512 0.000 0.488 0.000
#> SRR1274543 2 0.3919 0.5212 0.000 0.776 0.000 0.036 0.188
#> SRR1274544 2 0.4305 0.5713 0.000 0.512 0.000 0.488 0.000
#> SRR1274545 2 0.0510 0.6140 0.000 0.984 0.000 0.016 0.000
#> SRR1274546 2 0.4084 0.5997 0.000 0.668 0.000 0.328 0.004
#> SRR1274547 2 0.1117 0.6055 0.000 0.964 0.000 0.020 0.016
#> SRR1274548 2 0.4744 0.1991 0.000 0.572 0.000 0.020 0.408
#> SRR1274549 5 0.7694 -0.3609 0.120 0.044 0.076 0.204 0.556
#> SRR1274550 3 0.1704 0.7795 0.000 0.000 0.928 0.004 0.068
#> SRR1274551 2 0.4302 0.5735 0.000 0.520 0.000 0.480 0.000
#> SRR1274552 5 0.5909 0.0146 0.112 0.000 0.156 0.052 0.680
#> SRR1274553 5 0.6667 -0.1252 0.120 0.000 0.264 0.048 0.568
#> SRR1274554 5 0.2358 0.5828 0.000 0.104 0.000 0.008 0.888
#> SRR1274555 3 0.1197 0.7880 0.000 0.000 0.952 0.000 0.048
#> SRR1274556 3 0.0579 0.7747 0.000 0.000 0.984 0.008 0.008
#> SRR1274557 3 0.5877 0.3583 0.108 0.000 0.596 0.008 0.288
#> SRR1274558 2 0.4744 0.1676 0.000 0.572 0.000 0.020 0.408
#> SRR1274559 5 0.3375 0.5729 0.000 0.104 0.000 0.056 0.840
#> SRR1274560 5 0.5237 -0.1542 0.012 0.004 0.024 0.368 0.592
#> SRR1274561 4 0.7607 0.6399 0.120 0.000 0.104 0.396 0.380
#> SRR1274562 3 0.7221 -0.1007 0.080 0.000 0.504 0.296 0.120
#> SRR1274563 3 0.1357 0.7877 0.000 0.000 0.948 0.004 0.048
#> SRR1274564 2 0.6796 -0.1217 0.000 0.456 0.008 0.220 0.316
#> SRR1274565 2 0.4425 0.4612 0.000 0.716 0.000 0.040 0.244
#> SRR1274566 5 0.6865 0.1435 0.000 0.360 0.008 0.220 0.412
#> SRR1274567 5 0.6691 0.1341 0.004 0.180 0.016 0.256 0.544
#> SRR1274568 5 0.3934 0.4781 0.000 0.244 0.000 0.016 0.740
#> SRR1274569 4 0.8234 0.6741 0.232 0.000 0.152 0.396 0.220
#> SRR1274570 1 0.5097 0.5853 0.708 0.000 0.060 0.212 0.020
#> SRR1274571 1 0.5893 0.4838 0.660 0.000 0.072 0.216 0.052
#> SRR1274572 1 0.1197 0.8058 0.952 0.000 0.048 0.000 0.000
#> SRR1274573 2 0.4305 0.5713 0.000 0.512 0.000 0.488 0.000
#> SRR1274574 2 0.4410 0.5818 0.000 0.556 0.000 0.440 0.004
#> SRR1274575 2 0.2221 0.5782 0.000 0.912 0.000 0.036 0.052
#> SRR1274576 2 0.4238 0.2812 0.000 0.628 0.000 0.004 0.368
#> SRR1274577 5 0.2463 0.5841 0.000 0.100 0.008 0.004 0.888
#> SRR1274578 1 0.0000 0.8243 1.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.8243 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.8243 1.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.1924 0.7541 0.004 0.000 0.924 0.064 0.008
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.0260 0.912 0.000 0.992 0.000 0.000 0.000 0.008
#> SRR1274529 2 0.0458 0.907 0.000 0.984 0.000 0.000 0.000 0.016
#> SRR1274530 2 0.0405 0.915 0.000 0.988 0.000 0.000 0.008 0.004
#> SRR1274531 4 0.3816 0.671 0.000 0.016 0.000 0.688 0.296 0.000
#> SRR1274532 6 0.3023 0.949 0.000 0.232 0.000 0.000 0.000 0.768
#> SRR1274533 5 0.1967 0.903 0.000 0.012 0.000 0.084 0.904 0.000
#> SRR1274534 5 0.0458 0.869 0.000 0.016 0.000 0.000 0.984 0.000
#> SRR1274535 5 0.0984 0.875 0.000 0.012 0.008 0.012 0.968 0.000
#> SRR1274536 4 0.3259 0.723 0.000 0.012 0.000 0.772 0.216 0.000
#> SRR1274537 2 0.0508 0.910 0.000 0.984 0.000 0.000 0.004 0.012
#> SRR1274538 2 0.0291 0.914 0.000 0.992 0.000 0.004 0.004 0.000
#> SRR1274539 5 0.0976 0.870 0.000 0.016 0.008 0.008 0.968 0.000
#> SRR1274540 6 0.3023 0.949 0.000 0.232 0.000 0.000 0.000 0.768
#> SRR1274541 6 0.3076 0.946 0.000 0.240 0.000 0.000 0.000 0.760
#> SRR1274542 6 0.3023 0.949 0.000 0.232 0.000 0.000 0.000 0.768
#> SRR1274543 2 0.0146 0.915 0.000 0.996 0.000 0.000 0.004 0.000
#> SRR1274544 6 0.3023 0.949 0.000 0.232 0.000 0.000 0.000 0.768
#> SRR1274545 2 0.0458 0.907 0.000 0.984 0.000 0.000 0.000 0.016
#> SRR1274546 6 0.3717 0.770 0.000 0.384 0.000 0.000 0.000 0.616
#> SRR1274547 2 0.0405 0.915 0.000 0.988 0.000 0.000 0.008 0.004
#> SRR1274548 2 0.1867 0.864 0.000 0.916 0.000 0.020 0.064 0.000
#> SRR1274549 4 0.1802 0.661 0.000 0.012 0.000 0.916 0.072 0.000
#> SRR1274550 3 0.0508 0.821 0.000 0.000 0.984 0.004 0.000 0.012
#> SRR1274551 6 0.3151 0.940 0.000 0.252 0.000 0.000 0.000 0.748
#> SRR1274552 3 0.5272 0.405 0.004 0.000 0.500 0.420 0.072 0.004
#> SRR1274553 3 0.5423 0.604 0.036 0.000 0.600 0.304 0.056 0.004
#> SRR1274554 5 0.1967 0.903 0.000 0.012 0.000 0.084 0.904 0.000
#> SRR1274555 3 0.0000 0.821 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274556 3 0.1218 0.814 0.000 0.000 0.956 0.004 0.012 0.028
#> SRR1274557 3 0.2595 0.776 0.000 0.000 0.836 0.160 0.000 0.004
#> SRR1274558 2 0.2948 0.724 0.000 0.804 0.000 0.008 0.188 0.000
#> SRR1274559 5 0.1967 0.903 0.000 0.012 0.000 0.084 0.904 0.000
#> SRR1274560 4 0.3521 0.704 0.000 0.000 0.004 0.724 0.268 0.004
#> SRR1274561 4 0.2089 0.644 0.004 0.000 0.012 0.908 0.072 0.004
#> SRR1274562 3 0.3017 0.769 0.052 0.000 0.840 0.108 0.000 0.000
#> SRR1274563 3 0.0000 0.821 0.000 0.000 1.000 0.000 0.000 0.000
#> SRR1274564 2 0.4942 0.404 0.000 0.652 0.000 0.192 0.156 0.000
#> SRR1274565 2 0.0777 0.908 0.000 0.972 0.000 0.004 0.024 0.000
#> SRR1274566 4 0.5898 0.312 0.000 0.380 0.000 0.416 0.204 0.000
#> SRR1274567 4 0.3348 0.723 0.000 0.016 0.000 0.768 0.216 0.000
#> SRR1274568 5 0.3780 0.642 0.000 0.020 0.004 0.248 0.728 0.000
#> SRR1274569 1 0.5105 0.362 0.496 0.000 0.048 0.444 0.008 0.004
#> SRR1274570 1 0.1124 0.862 0.956 0.000 0.008 0.036 0.000 0.000
#> SRR1274571 1 0.3507 0.732 0.752 0.000 0.012 0.232 0.000 0.004
#> SRR1274572 1 0.0146 0.873 0.996 0.000 0.000 0.004 0.000 0.000
#> SRR1274573 6 0.3023 0.949 0.000 0.232 0.000 0.000 0.000 0.768
#> SRR1274574 6 0.3659 0.809 0.000 0.364 0.000 0.000 0.000 0.636
#> SRR1274575 2 0.0405 0.915 0.000 0.988 0.000 0.000 0.008 0.004
#> SRR1274576 2 0.1082 0.896 0.000 0.956 0.000 0.004 0.040 0.000
#> SRR1274577 5 0.1967 0.903 0.000 0.012 0.000 0.084 0.904 0.000
#> SRR1274578 1 0.0000 0.873 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274579 1 0.0000 0.873 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0000 0.873 1.000 0.000 0.000 0.000 0.000 0.000
#> SRR1274581 3 0.3342 0.734 0.000 0.000 0.760 0.000 0.012 0.228
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
The object with results only for a single top-value method and a single partition method can be extracted as:
res = res_list["ATC", "NMF"]
# you can also extract it by
# res = res_list["ATC:NMF"]
A summary of res and all the functions that can be applied to it:
res
#> A 'ConsensusPartition' object with k = 2, 3, 4, 5, 6.
#> On a matrix with 16571 rows and 54 columns.
#> Top rows (1000, 2000, 3000, 4000, 5000) are extracted by 'ATC' method.
#> Subgroups are detected by 'NMF' method.
#> Performed in total 1250 partitions by row resampling.
#> Best k for subgroups seems to be 2.
#>
#> Following methods can be applied to this 'ConsensusPartition' object:
#> [1] "cola_report" "collect_classes" "collect_plots"
#> [4] "collect_stats" "colnames" "compare_signatures"
#> [7] "consensus_heatmap" "dimension_reduction" "functional_enrichment"
#> [10] "get_anno_col" "get_anno" "get_classes"
#> [13] "get_consensus" "get_matrix" "get_membership"
#> [16] "get_param" "get_signatures" "get_stats"
#> [19] "is_best_k" "is_stable_k" "membership_heatmap"
#> [22] "ncol" "nrow" "plot_ecdf"
#> [25] "rownames" "select_partition_number" "show"
#> [28] "suggest_best_k" "test_to_known_factors"
collect_plots() function collects all the plots made from res for all k (number of partitions)
into one single page to provide an easy and fast comparison between different k.
collect_plots(res)
The plots are:
k and the heatmap of
predicted classes for each k.k.k.k.All the plots in panels can be made by individual functions and they are plotted later in this section.
select_partition_number() produces several plots showing different
statistics for choosing “optimized” k. There are following statistics:
k;k, the area increased is defined as \(A_k - A_{k-1}\).The detailed explanations of these statistics can be found in the cola vignette.
Generally speaking, lower PAC score, higher mean silhouette score or higher
concordance corresponds to better partition. Rand index and Jaccard index
measure how similar the current partition is compared to partition with k-1.
If they are too similar, we won't accept k is better than k-1.
select_partition_number(res)
The numeric values for all these statistics can be obtained by get_stats().
get_stats(res)
#> k 1-PAC mean_silhouette concordance area_increased Rand Jaccard
#> 2 2 1.000 0.982 0.992 0.4626 0.535 0.535
#> 3 3 0.638 0.702 0.849 0.1773 0.913 0.844
#> 4 4 0.599 0.790 0.837 0.2642 0.732 0.493
#> 5 5 0.653 0.700 0.841 0.0641 0.993 0.976
#> 6 6 0.647 0.626 0.815 0.0411 0.844 0.549
suggest_best_k() suggests the best \(k\) based on these statistics. The rules are as follows:
suggest_best_k(res)
#> [1] 2
Following shows the table of the partitions (You need to click the show/hide
code output link to see it). The membership matrix (columns with name p*)
is inferred by
clue::cl_consensus()
function with the SE method. Basically the value in the membership matrix
represents the probability to belong to a certain group. The finall class
label for an item is determined with the group with highest probability it
belongs to.
In get_classes() function, the entropy is calculated from the membership
matrix and the silhouette score is calculated from the consensus matrix.
cbind(get_classes(res, k = 2), get_membership(res, k = 2))
#> class entropy silhouette p1 p2
#> SRR1274528 2 0.000 0.997 0.000 1.000
#> SRR1274529 2 0.000 0.997 0.000 1.000
#> SRR1274530 2 0.000 0.997 0.000 1.000
#> SRR1274531 2 0.295 0.946 0.052 0.948
#> SRR1274532 2 0.000 0.997 0.000 1.000
#> SRR1274533 2 0.000 0.997 0.000 1.000
#> SRR1274534 2 0.000 0.997 0.000 1.000
#> SRR1274535 2 0.278 0.950 0.048 0.952
#> SRR1274536 2 0.000 0.997 0.000 1.000
#> SRR1274537 2 0.000 0.997 0.000 1.000
#> SRR1274538 2 0.000 0.997 0.000 1.000
#> SRR1274539 2 0.000 0.997 0.000 1.000
#> SRR1274540 2 0.000 0.997 0.000 1.000
#> SRR1274541 2 0.000 0.997 0.000 1.000
#> SRR1274542 2 0.000 0.997 0.000 1.000
#> SRR1274543 2 0.000 0.997 0.000 1.000
#> SRR1274544 2 0.000 0.997 0.000 1.000
#> SRR1274545 2 0.000 0.997 0.000 1.000
#> SRR1274546 2 0.000 0.997 0.000 1.000
#> SRR1274547 2 0.000 0.997 0.000 1.000
#> SRR1274548 2 0.000 0.997 0.000 1.000
#> SRR1274549 1 0.000 0.983 1.000 0.000
#> SRR1274550 1 0.000 0.983 1.000 0.000
#> SRR1274551 2 0.000 0.997 0.000 1.000
#> SRR1274552 1 0.000 0.983 1.000 0.000
#> SRR1274553 1 0.000 0.983 1.000 0.000
#> SRR1274554 2 0.000 0.997 0.000 1.000
#> SRR1274555 1 0.000 0.983 1.000 0.000
#> SRR1274556 1 0.000 0.983 1.000 0.000
#> SRR1274557 1 0.000 0.983 1.000 0.000
#> SRR1274558 2 0.000 0.997 0.000 1.000
#> SRR1274559 2 0.000 0.997 0.000 1.000
#> SRR1274560 1 0.891 0.551 0.692 0.308
#> SRR1274561 1 0.000 0.983 1.000 0.000
#> SRR1274562 1 0.000 0.983 1.000 0.000
#> SRR1274563 1 0.000 0.983 1.000 0.000
#> SRR1274564 2 0.000 0.997 0.000 1.000
#> SRR1274565 2 0.000 0.997 0.000 1.000
#> SRR1274566 2 0.000 0.997 0.000 1.000
#> SRR1274567 2 0.000 0.997 0.000 1.000
#> SRR1274568 2 0.000 0.997 0.000 1.000
#> SRR1274569 1 0.000 0.983 1.000 0.000
#> SRR1274570 1 0.000 0.983 1.000 0.000
#> SRR1274571 1 0.000 0.983 1.000 0.000
#> SRR1274572 1 0.000 0.983 1.000 0.000
#> SRR1274573 2 0.000 0.997 0.000 1.000
#> SRR1274574 2 0.000 0.997 0.000 1.000
#> SRR1274575 2 0.000 0.997 0.000 1.000
#> SRR1274576 2 0.000 0.997 0.000 1.000
#> SRR1274577 2 0.000 0.997 0.000 1.000
#> SRR1274578 1 0.000 0.983 1.000 0.000
#> SRR1274579 1 0.000 0.983 1.000 0.000
#> SRR1274580 1 0.000 0.983 1.000 0.000
#> SRR1274581 1 0.000 0.983 1.000 0.000
cbind(get_classes(res, k = 3), get_membership(res, k = 3))
#> class entropy silhouette p1 p2 p3
#> SRR1274528 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274529 2 0.3551 0.877 0.132 0.868 0.000
#> SRR1274530 2 0.3412 0.878 0.124 0.876 0.000
#> SRR1274531 2 0.8573 0.457 0.116 0.556 0.328
#> SRR1274532 2 0.0237 0.917 0.004 0.996 0.000
#> SRR1274533 2 0.0424 0.916 0.008 0.992 0.000
#> SRR1274534 2 0.0475 0.916 0.004 0.992 0.004
#> SRR1274535 2 0.5882 0.509 0.000 0.652 0.348
#> SRR1274536 2 0.3826 0.875 0.124 0.868 0.008
#> SRR1274537 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274538 2 0.0424 0.917 0.008 0.992 0.000
#> SRR1274539 2 0.0747 0.913 0.000 0.984 0.016
#> SRR1274540 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274541 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274542 2 0.1289 0.908 0.032 0.968 0.000
#> SRR1274543 2 0.0237 0.917 0.004 0.996 0.000
#> SRR1274544 2 0.0592 0.915 0.012 0.988 0.000
#> SRR1274545 2 0.3038 0.889 0.104 0.896 0.000
#> SRR1274546 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274547 2 0.2796 0.892 0.092 0.908 0.000
#> SRR1274548 2 0.0424 0.917 0.008 0.992 0.000
#> SRR1274549 3 0.3267 0.545 0.116 0.000 0.884
#> SRR1274550 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274551 2 0.0237 0.917 0.004 0.996 0.000
#> SRR1274552 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274553 3 0.0892 0.677 0.020 0.000 0.980
#> SRR1274554 2 0.4235 0.802 0.176 0.824 0.000
#> SRR1274555 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274556 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274557 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274558 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274559 2 0.5012 0.770 0.204 0.788 0.008
#> SRR1274560 2 0.9953 -0.370 0.344 0.368 0.288
#> SRR1274561 3 0.6500 -0.589 0.464 0.004 0.532
#> SRR1274562 3 0.5810 -0.137 0.336 0.000 0.664
#> SRR1274563 3 0.0000 0.692 0.000 0.000 1.000
#> SRR1274564 2 0.3619 0.874 0.136 0.864 0.000
#> SRR1274565 2 0.2711 0.899 0.088 0.912 0.000
#> SRR1274566 2 0.3482 0.877 0.128 0.872 0.000
#> SRR1274567 2 0.3482 0.877 0.128 0.872 0.000
#> SRR1274568 2 0.0000 0.917 0.000 1.000 0.000
#> SRR1274569 3 0.6305 -0.639 0.484 0.000 0.516
#> SRR1274570 1 0.6244 0.761 0.560 0.000 0.440
#> SRR1274571 1 0.6274 0.730 0.544 0.000 0.456
#> SRR1274572 1 0.6244 0.761 0.560 0.000 0.440
#> SRR1274573 2 0.0592 0.915 0.012 0.988 0.000
#> SRR1274574 2 0.0592 0.916 0.012 0.988 0.000
#> SRR1274575 2 0.3412 0.878 0.124 0.876 0.000
#> SRR1274576 2 0.0237 0.916 0.004 0.996 0.000
#> SRR1274577 2 0.0424 0.916 0.008 0.992 0.000
#> SRR1274578 1 0.3619 0.506 0.864 0.000 0.136
#> SRR1274579 1 0.5706 0.717 0.680 0.000 0.320
#> SRR1274580 1 0.6126 0.769 0.600 0.000 0.400
#> SRR1274581 3 0.6267 -0.540 0.452 0.000 0.548
cbind(get_classes(res, k = 4), get_membership(res, k = 4))
#> class entropy silhouette p1 p2 p3 p4
#> SRR1274528 2 0.1022 0.863 0.000 0.968 0.000 0.032
#> SRR1274529 4 0.4164 0.821 0.000 0.264 0.000 0.736
#> SRR1274530 4 0.3688 0.851 0.000 0.208 0.000 0.792
#> SRR1274531 4 0.7486 0.486 0.000 0.188 0.348 0.464
#> SRR1274532 2 0.1706 0.862 0.016 0.948 0.000 0.036
#> SRR1274533 2 0.2867 0.814 0.012 0.884 0.000 0.104
#> SRR1274534 2 0.2654 0.815 0.004 0.888 0.000 0.108
#> SRR1274535 2 0.6168 0.326 0.000 0.556 0.388 0.056
#> SRR1274536 4 0.3626 0.850 0.004 0.184 0.000 0.812
#> SRR1274537 2 0.1661 0.847 0.004 0.944 0.000 0.052
#> SRR1274538 2 0.2011 0.836 0.000 0.920 0.000 0.080
#> SRR1274539 2 0.2805 0.819 0.000 0.888 0.012 0.100
#> SRR1274540 2 0.1890 0.853 0.008 0.936 0.000 0.056
#> SRR1274541 2 0.1118 0.862 0.000 0.964 0.000 0.036
#> SRR1274542 2 0.2494 0.853 0.048 0.916 0.000 0.036
#> SRR1274543 2 0.1302 0.860 0.000 0.956 0.000 0.044
#> SRR1274544 2 0.1833 0.862 0.024 0.944 0.000 0.032
#> SRR1274545 4 0.4624 0.747 0.000 0.340 0.000 0.660
#> SRR1274546 2 0.0336 0.865 0.000 0.992 0.000 0.008
#> SRR1274547 4 0.4761 0.700 0.000 0.372 0.000 0.628
#> SRR1274548 2 0.1716 0.850 0.000 0.936 0.000 0.064
#> SRR1274549 4 0.5510 0.465 0.008 0.032 0.276 0.684
#> SRR1274550 3 0.0000 0.981 0.000 0.000 1.000 0.000
#> SRR1274551 2 0.2125 0.839 0.004 0.920 0.000 0.076
#> SRR1274552 3 0.0927 0.967 0.008 0.000 0.976 0.016
#> SRR1274553 3 0.2060 0.921 0.052 0.000 0.932 0.016
#> SRR1274554 2 0.4610 0.748 0.100 0.800 0.000 0.100
#> SRR1274555 3 0.0188 0.980 0.000 0.000 0.996 0.004
#> SRR1274556 3 0.0188 0.980 0.000 0.000 0.996 0.004
#> SRR1274557 3 0.0000 0.981 0.000 0.000 1.000 0.000
#> SRR1274558 2 0.0000 0.864 0.000 1.000 0.000 0.000
#> SRR1274559 2 0.7383 0.123 0.424 0.464 0.024 0.088
#> SRR1274560 1 0.7810 0.603 0.576 0.164 0.216 0.044
#> SRR1274561 1 0.6227 0.756 0.672 0.004 0.212 0.112
#> SRR1274562 1 0.4998 0.497 0.512 0.000 0.488 0.000
#> SRR1274563 3 0.0000 0.981 0.000 0.000 1.000 0.000
#> SRR1274564 4 0.3626 0.850 0.004 0.184 0.000 0.812
#> SRR1274565 4 0.4991 0.667 0.004 0.388 0.000 0.608
#> SRR1274566 4 0.3710 0.852 0.004 0.192 0.000 0.804
#> SRR1274567 4 0.3626 0.850 0.004 0.184 0.000 0.812
#> SRR1274568 2 0.1022 0.861 0.000 0.968 0.000 0.032
#> SRR1274569 1 0.4511 0.798 0.724 0.000 0.268 0.008
#> SRR1274570 1 0.3837 0.817 0.776 0.000 0.224 0.000
#> SRR1274571 1 0.3801 0.818 0.780 0.000 0.220 0.000
#> SRR1274572 1 0.3945 0.819 0.780 0.000 0.216 0.004
#> SRR1274573 2 0.1724 0.862 0.020 0.948 0.000 0.032
#> SRR1274574 2 0.3428 0.746 0.012 0.844 0.000 0.144
#> SRR1274575 4 0.3688 0.851 0.000 0.208 0.000 0.792
#> SRR1274576 2 0.2081 0.832 0.000 0.916 0.000 0.084
#> SRR1274577 2 0.2611 0.822 0.008 0.896 0.000 0.096
#> SRR1274578 1 0.0657 0.705 0.984 0.004 0.000 0.012
#> SRR1274579 1 0.2081 0.772 0.916 0.000 0.084 0.000
#> SRR1274580 1 0.3219 0.810 0.836 0.000 0.164 0.000
#> SRR1274581 1 0.6477 0.585 0.552 0.000 0.368 0.080
cbind(get_classes(res, k = 5), get_membership(res, k = 5))
#> class entropy silhouette p1 p2 p3 p4 p5
#> SRR1274528 2 0.1750 0.831 0.000 0.936 0.000 0.036 0.028
#> SRR1274529 4 0.4473 0.604 0.000 0.324 0.000 0.656 0.020
#> SRR1274530 4 0.1768 0.727 0.000 0.072 0.000 0.924 0.004
#> SRR1274531 4 0.5104 0.374 0.000 0.044 0.304 0.644 0.008
#> SRR1274532 2 0.2122 0.823 0.008 0.924 0.000 0.032 0.036
#> SRR1274533 2 0.2818 0.793 0.000 0.856 0.012 0.000 0.132
#> SRR1274534 2 0.3885 0.689 0.000 0.724 0.008 0.000 0.268
#> SRR1274535 2 0.6597 0.225 0.000 0.460 0.296 0.000 0.244
#> SRR1274536 4 0.0955 0.716 0.004 0.028 0.000 0.968 0.000
#> SRR1274537 2 0.0703 0.833 0.000 0.976 0.000 0.000 0.024
#> SRR1274538 2 0.1792 0.806 0.000 0.916 0.000 0.084 0.000
#> SRR1274539 2 0.4790 0.645 0.000 0.676 0.032 0.008 0.284
#> SRR1274540 2 0.2520 0.808 0.000 0.896 0.000 0.056 0.048
#> SRR1274541 2 0.1082 0.831 0.000 0.964 0.000 0.028 0.008
#> SRR1274542 2 0.1740 0.824 0.012 0.932 0.000 0.000 0.056
#> SRR1274543 2 0.1197 0.825 0.000 0.952 0.000 0.048 0.000
#> SRR1274544 2 0.1605 0.828 0.012 0.944 0.000 0.004 0.040
#> SRR1274545 4 0.5065 0.430 0.000 0.420 0.000 0.544 0.036
#> SRR1274546 2 0.0162 0.834 0.000 0.996 0.000 0.000 0.004
#> SRR1274547 4 0.4192 0.487 0.000 0.404 0.000 0.596 0.000
#> SRR1274548 2 0.3346 0.804 0.000 0.844 0.000 0.064 0.092
#> SRR1274549 4 0.3066 0.596 0.008 0.012 0.116 0.860 0.004
#> SRR1274550 3 0.1792 0.836 0.000 0.000 0.916 0.000 0.084
#> SRR1274551 2 0.2172 0.806 0.000 0.908 0.000 0.076 0.016
#> SRR1274552 3 0.3446 0.749 0.004 0.004 0.812 0.008 0.172
#> SRR1274553 3 0.3936 0.744 0.052 0.000 0.800 0.004 0.144
#> SRR1274554 2 0.2529 0.818 0.040 0.900 0.004 0.000 0.056
#> SRR1274555 3 0.0324 0.860 0.004 0.000 0.992 0.000 0.004
#> SRR1274556 3 0.1364 0.829 0.012 0.000 0.952 0.000 0.036
#> SRR1274557 3 0.0451 0.859 0.004 0.000 0.988 0.000 0.008
#> SRR1274558 2 0.0404 0.834 0.000 0.988 0.000 0.000 0.012
#> SRR1274559 2 0.7472 0.223 0.176 0.500 0.008 0.056 0.260
#> SRR1274560 1 0.5800 0.591 0.708 0.064 0.052 0.160 0.016
#> SRR1274561 1 0.5335 0.552 0.676 0.000 0.112 0.208 0.004
#> SRR1274562 1 0.4446 0.327 0.592 0.000 0.400 0.000 0.008
#> SRR1274563 3 0.0579 0.857 0.008 0.000 0.984 0.000 0.008
#> SRR1274564 4 0.1197 0.727 0.000 0.048 0.000 0.952 0.000
#> SRR1274565 4 0.5005 0.497 0.004 0.388 0.000 0.580 0.028
#> SRR1274566 4 0.0880 0.720 0.000 0.032 0.000 0.968 0.000
#> SRR1274567 4 0.1369 0.711 0.008 0.028 0.000 0.956 0.008
#> SRR1274568 2 0.4765 0.697 0.000 0.704 0.000 0.068 0.228
#> SRR1274569 1 0.3619 0.774 0.844 0.000 0.092 0.036 0.028
#> SRR1274570 1 0.1952 0.797 0.912 0.000 0.084 0.004 0.000
#> SRR1274571 1 0.2112 0.796 0.908 0.000 0.084 0.004 0.004
#> SRR1274572 1 0.1544 0.799 0.932 0.000 0.068 0.000 0.000
#> SRR1274573 2 0.0968 0.834 0.012 0.972 0.000 0.012 0.004
#> SRR1274574 2 0.3205 0.693 0.004 0.816 0.000 0.176 0.004
#> SRR1274575 4 0.1410 0.729 0.000 0.060 0.000 0.940 0.000
#> SRR1274576 2 0.2020 0.813 0.000 0.900 0.000 0.000 0.100
#> SRR1274577 2 0.3968 0.681 0.004 0.716 0.004 0.000 0.276
#> SRR1274578 1 0.0703 0.765 0.976 0.000 0.000 0.000 0.024
#> SRR1274579 1 0.0000 0.776 1.000 0.000 0.000 0.000 0.000
#> SRR1274580 1 0.0510 0.786 0.984 0.000 0.016 0.000 0.000
#> SRR1274581 5 0.7136 0.000 0.232 0.000 0.296 0.024 0.448
cbind(get_classes(res, k = 6), get_membership(res, k = 6))
#> class entropy silhouette p1 p2 p3 p4 p5 p6
#> SRR1274528 2 0.2830 0.6959 0.000 0.836 0.000 0.020 0.144 0.000
#> SRR1274529 2 0.4960 0.3401 0.000 0.568 0.000 0.376 0.032 0.024
#> SRR1274530 4 0.2400 0.8337 0.000 0.116 0.000 0.872 0.008 0.004
#> SRR1274531 4 0.4818 0.5462 0.000 0.072 0.272 0.648 0.008 0.000
#> SRR1274532 2 0.0790 0.7648 0.000 0.968 0.000 0.000 0.032 0.000
#> SRR1274533 2 0.4483 0.0687 0.000 0.548 0.016 0.004 0.428 0.004
#> SRR1274534 5 0.3672 0.5801 0.000 0.304 0.008 0.000 0.688 0.000
#> SRR1274535 5 0.5032 0.4444 0.000 0.132 0.216 0.000 0.648 0.004
#> SRR1274536 4 0.1296 0.8969 0.004 0.044 0.000 0.948 0.004 0.000
#> SRR1274537 2 0.2416 0.6840 0.000 0.844 0.000 0.000 0.156 0.000
#> SRR1274538 2 0.1320 0.7685 0.000 0.948 0.000 0.036 0.016 0.000
#> SRR1274539 5 0.3895 0.5834 0.000 0.284 0.016 0.004 0.696 0.000
#> SRR1274540 2 0.1313 0.7661 0.000 0.952 0.000 0.016 0.028 0.004
#> SRR1274541 2 0.0508 0.7691 0.000 0.984 0.000 0.012 0.004 0.000
#> SRR1274542 2 0.1881 0.7500 0.008 0.928 0.000 0.012 0.044 0.008
#> SRR1274543 2 0.1564 0.7628 0.000 0.936 0.000 0.024 0.040 0.000
#> SRR1274544 2 0.1194 0.7607 0.004 0.956 0.000 0.008 0.032 0.000
#> SRR1274545 2 0.4514 0.5559 0.000 0.680 0.000 0.264 0.040 0.016
#> SRR1274546 2 0.0972 0.7660 0.000 0.964 0.000 0.008 0.028 0.000
#> SRR1274547 2 0.3945 0.4340 0.000 0.612 0.000 0.380 0.008 0.000
#> SRR1274548 2 0.3993 0.4992 0.000 0.700 0.000 0.024 0.272 0.004
#> SRR1274549 4 0.2372 0.8491 0.008 0.024 0.036 0.908 0.024 0.000
#> SRR1274550 3 0.3383 0.5548 0.000 0.000 0.728 0.004 0.268 0.000
#> SRR1274551 2 0.0790 0.7695 0.000 0.968 0.000 0.032 0.000 0.000
#> SRR1274552 5 0.4491 -0.2367 0.000 0.000 0.476 0.008 0.500 0.016
#> SRR1274553 5 0.5968 -0.2568 0.084 0.000 0.440 0.008 0.440 0.028
#> SRR1274554 2 0.4119 0.6727 0.016 0.800 0.008 0.028 0.120 0.028
#> SRR1274555 3 0.0363 0.8645 0.012 0.000 0.988 0.000 0.000 0.000
#> SRR1274556 3 0.1780 0.8325 0.024 0.000 0.932 0.004 0.004 0.036
#> SRR1274557 3 0.0291 0.8620 0.004 0.000 0.992 0.000 0.004 0.000
#> SRR1274558 2 0.1812 0.7445 0.000 0.912 0.000 0.008 0.080 0.000
#> SRR1274559 2 0.6824 0.4550 0.032 0.616 0.048 0.064 0.164 0.076
#> SRR1274560 1 0.3900 0.6835 0.804 0.056 0.020 0.112 0.008 0.000
#> SRR1274561 1 0.5695 0.3625 0.544 0.000 0.260 0.192 0.000 0.004
#> SRR1274562 1 0.3864 0.2312 0.520 0.000 0.480 0.000 0.000 0.000
#> SRR1274563 3 0.1196 0.8499 0.040 0.000 0.952 0.000 0.008 0.000
#> SRR1274564 4 0.2002 0.8921 0.004 0.052 0.000 0.920 0.012 0.012
#> SRR1274565 2 0.4883 0.3890 0.000 0.588 0.000 0.356 0.040 0.016
#> SRR1274566 4 0.1219 0.8974 0.004 0.048 0.000 0.948 0.000 0.000
#> SRR1274567 4 0.1523 0.8940 0.008 0.044 0.000 0.940 0.008 0.000
#> SRR1274568 5 0.5003 0.2320 0.004 0.440 0.000 0.048 0.504 0.004
#> SRR1274569 1 0.2078 0.7927 0.912 0.000 0.040 0.044 0.004 0.000
#> SRR1274570 1 0.0837 0.8118 0.972 0.000 0.020 0.004 0.000 0.004
#> SRR1274571 1 0.0922 0.8115 0.968 0.000 0.024 0.004 0.000 0.004
#> SRR1274572 1 0.0951 0.8121 0.968 0.000 0.020 0.000 0.004 0.008
#> SRR1274573 2 0.0951 0.7675 0.008 0.968 0.000 0.004 0.020 0.000
#> SRR1274574 2 0.1643 0.7614 0.000 0.924 0.000 0.068 0.008 0.000
#> SRR1274575 4 0.1700 0.8792 0.000 0.080 0.000 0.916 0.004 0.000
#> SRR1274576 2 0.3827 0.4434 0.000 0.680 0.000 0.008 0.308 0.004
#> SRR1274577 5 0.3894 0.5344 0.004 0.324 0.000 0.000 0.664 0.008
#> SRR1274578 1 0.2180 0.7712 0.912 0.004 0.008 0.000 0.048 0.028
#> SRR1274579 1 0.1176 0.7969 0.956 0.000 0.000 0.000 0.024 0.020
#> SRR1274580 1 0.1078 0.8061 0.964 0.000 0.008 0.000 0.016 0.012
#> SRR1274581 6 0.3206 0.0000 0.068 0.000 0.104 0.000 0.000 0.828
Heatmaps for the consensus matrix. It visualizes the probability of two samples to be in a same group.
consensus_heatmap(res, k = 2)
consensus_heatmap(res, k = 3)
consensus_heatmap(res, k = 4)
consensus_heatmap(res, k = 5)
consensus_heatmap(res, k = 6)
Heatmaps for the membership of samples in all partitions to see how consistent they are:
membership_heatmap(res, k = 2)
membership_heatmap(res, k = 3)
membership_heatmap(res, k = 4)
membership_heatmap(res, k = 5)
membership_heatmap(res, k = 6)
As soon as we have had the classes for columns, we can look for signatures which are significantly different between classes which can be candidate marks for certain classes. Following are the heatmaps for signatures.
Signature heatmaps where rows are scaled:
get_signatures(res, k = 2)
get_signatures(res, k = 3)
get_signatures(res, k = 4)
get_signatures(res, k = 5)
get_signatures(res, k = 6)
Signature heatmaps where rows are not scaled:
get_signatures(res, k = 2, scale_rows = FALSE)
get_signatures(res, k = 3, scale_rows = FALSE)
get_signatures(res, k = 4, scale_rows = FALSE)
get_signatures(res, k = 5, scale_rows = FALSE)
get_signatures(res, k = 6, scale_rows = FALSE)
Compare the overlap of signatures from different k:
compare_signatures(res)
get_signature() returns a data frame invisibly. TO get the list of signatures, the function
call should be assigned to a variable explicitly. In following code, if plot argument is set
to FALSE, no heatmap is plotted while only the differential analysis is performed.
# code only for demonstration
tb = get_signature(res, k = ..., plot = FALSE)
An example of the output of tb is:
#> which_row fdr mean_1 mean_2 scaled_mean_1 scaled_mean_2 km
#> 1 38 0.042760348 8.373488 9.131774 -0.5533452 0.5164555 1
#> 2 40 0.018707592 7.106213 8.469186 -0.6173731 0.5762149 1
#> 3 55 0.019134737 10.221463 11.207825 -0.6159697 0.5749050 1
#> 4 59 0.006059896 5.921854 7.869574 -0.6899429 0.6439467 1
#> 5 60 0.018055526 8.928898 10.211722 -0.6204761 0.5791110 1
#> 6 98 0.009384629 15.714769 14.887706 0.6635654 -0.6193277 2
...
The columns in tb are:
which_row: row indices corresponding to the input matrix.fdr: FDR for the differential test. mean_x: The mean value in group x.scaled_mean_x: The mean value in group x after rows are scaled.km: Row groups if k-means clustering is applied to rows.UMAP plot which shows how samples are separated.
dimension_reduction(res, k = 2, method = "UMAP")
dimension_reduction(res, k = 3, method = "UMAP")
dimension_reduction(res, k = 4, method = "UMAP")
dimension_reduction(res, k = 5, method = "UMAP")
dimension_reduction(res, k = 6, method = "UMAP")
Following heatmap shows how subgroups are split when increasing k:
collect_classes(res)
If matrix rows can be associated to genes, consider to use functional_enrichment(res,
...) to perform function enrichment for the signature genes. See this vignette for more detailed explanations.
sessionInfo()
#> R version 3.6.0 (2019-04-26)
#> Platform: x86_64-pc-linux-gnu (64-bit)
#> Running under: CentOS Linux 7 (Core)
#>
#> Matrix products: default
#> BLAS: /usr/lib64/libblas.so.3.4.2
#> LAPACK: /usr/lib64/liblapack.so.3.4.2
#>
#> locale:
#> [1] LC_CTYPE=en_GB.UTF-8 LC_NUMERIC=C LC_TIME=en_GB.UTF-8
#> [4] LC_COLLATE=en_GB.UTF-8 LC_MONETARY=en_GB.UTF-8 LC_MESSAGES=en_GB.UTF-8
#> [7] LC_PAPER=en_GB.UTF-8 LC_NAME=C LC_ADDRESS=C
#> [10] LC_TELEPHONE=C LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C
#>
#> attached base packages:
#> [1] grid stats graphics grDevices utils datasets methods base
#>
#> other attached packages:
#> [1] genefilter_1.66.0 ComplexHeatmap_2.3.1 markdown_1.1 knitr_1.26
#> [5] GetoptLong_0.1.7 cola_1.3.2
#>
#> loaded via a namespace (and not attached):
#> [1] circlize_0.4.8 shape_1.4.4 xfun_0.11 slam_0.1-46
#> [5] lattice_0.20-38 splines_3.6.0 colorspace_1.4-1 vctrs_0.2.0
#> [9] stats4_3.6.0 blob_1.2.0 XML_3.98-1.20 survival_2.44-1.1
#> [13] rlang_0.4.2 pillar_1.4.2 DBI_1.0.0 BiocGenerics_0.30.0
#> [17] bit64_0.9-7 RColorBrewer_1.1-2 matrixStats_0.55.0 stringr_1.4.0
#> [21] GlobalOptions_0.1.1 evaluate_0.14 memoise_1.1.0 Biobase_2.44.0
#> [25] IRanges_2.18.3 parallel_3.6.0 AnnotationDbi_1.46.1 highr_0.8
#> [29] Rcpp_1.0.3 xtable_1.8-4 backports_1.1.5 S4Vectors_0.22.1
#> [33] annotate_1.62.0 skmeans_0.2-11 bit_1.1-14 microbenchmark_1.4-7
#> [37] brew_1.0-6 impute_1.58.0 rjson_0.2.20 png_0.1-7
#> [41] digest_0.6.23 stringi_1.4.3 polyclip_1.10-0 clue_0.3-57
#> [45] tools_3.6.0 bitops_1.0-6 magrittr_1.5 eulerr_6.0.0
#> [49] RCurl_1.95-4.12 RSQLite_2.1.4 tibble_2.1.3 cluster_2.1.0
#> [53] crayon_1.3.4 pkgconfig_2.0.3 zeallot_0.1.0 Matrix_1.2-17
#> [57] xml2_1.2.2 httr_1.4.1 R6_2.4.1 mclust_5.4.5
#> [61] compiler_3.6.0