Correlation between aggregated molecular cancer subtypes and selected clinical features
Uterine Carcinosarcoma (Primary solid tumor)
15 July 2014  |  analyses__2014_07_15
Maintainer Information
Citation Information
Maintained by TCGA GDAC Team (Broad Institute/MD Anderson Cancer Center/Harvard Medical School)
Cite as Broad Institute TCGA Genome Data Analysis Center (2014): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1WS8S2H
Overview
Introduction

This pipeline computes the correlation between cancer subtypes identified by different molecular patterns and selected clinical features.

Summary

Testing the association between subtypes identified by 8 different clustering approaches and 3 clinical features across 57 patients, no significant finding detected with P value < 0.05 and Q value < 0.25.

  • 3 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes do not correlate to any clinical features.

  • 5 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes do not correlate to any clinical features.

  • CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that do not correlate to any clinical features.

  • Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 7 subtypes that do not correlate to any clinical features.

  • 4 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes do not correlate to any clinical features.

  • 4 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes do not correlate to any clinical features.

  • 4 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes do not correlate to any clinical features.

  • 6 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes do not correlate to any clinical features.

Results
Overview of the results

Table 1.  Get Full Table Overview of the association between subtypes identified by 8 different clustering approaches and 3 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, no significant finding detected.

Clinical
Features
Time
to
Death
AGE RACE
Statistical Tests logrank test Kruskal-Wallis (anova) Fisher's exact test
Copy Number Ratio CNMF subtypes 0.228
(1.00)
0.779
(1.00)
0.733
(1.00)
METHLYATION CNMF 0.885
(1.00)
0.102
(1.00)
0.955
(1.00)
RNAseq CNMF subtypes 0.99
(1.00)
0.205
(1.00)
0.658
(1.00)
RNAseq cHierClus subtypes 0.0971
(1.00)
0.234
(1.00)
0.845
(1.00)
MIRSEQ CNMF 0.11
(1.00)
0.496
(1.00)
0.828
(1.00)
MIRSEQ CHIERARCHICAL 0.935
(1.00)
0.209
(1.00)
0.81
(1.00)
MIRseq Mature CNMF subtypes 0.156
(1.00)
0.221
(1.00)
0.266
(1.00)
MIRseq Mature cHierClus subtypes 0.681
(1.00)
0.275
(1.00)
0.634
(1.00)
Clustering Approach #1: 'Copy Number Ratio CNMF subtypes'

Table S1.  Description of clustering approach #1: 'Copy Number Ratio CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 22 24 10
'Copy Number Ratio CNMF subtypes' versus 'Time to Death'

P value = 0.228 (logrank test), Q value = 1

Table S2.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 55 33 0.3 - 93.4 (18.1)
subtype1 22 13 6.7 - 93.4 (27.0)
subtype2 24 15 2.7 - 59.7 (15.3)
subtype3 9 5 0.3 - 85.3 (15.8)

Figure S1.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.779 (Kruskal-Wallis (anova)), Q value = 1

Table S3.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 56 69.8 (9.3)
subtype1 22 69.0 (8.1)
subtype2 24 70.8 (10.0)
subtype3 10 69.2 (10.8)

Figure S2.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.733 (Fisher's exact test), Q value = 1

Table S4.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 43
subtype1 1 5 15
subtype2 1 3 20
subtype3 1 1 8

Figure S3.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'RACE'

Clustering Approach #2: 'METHLYATION CNMF'

Table S5.  Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels 1 2 3 4 5
Number of samples 12 9 10 13 13
'METHLYATION CNMF' versus 'Time to Death'

P value = 0.885 (logrank test), Q value = 1

Table S6.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 56 34 0.3 - 102.4 (18.4)
subtype1 12 8 0.3 - 102.4 (27.0)
subtype2 9 6 4.7 - 69.6 (17.4)
subtype3 10 7 5.5 - 59.4 (18.8)
subtype4 12 6 2.7 - 93.4 (14.6)
subtype5 13 7 3.8 - 85.3 (19.6)

Figure S4.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'AGE'

P value = 0.102 (Kruskal-Wallis (anova)), Q value = 1

Table S7.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 57 69.7 (9.3)
subtype1 12 67.2 (8.0)
subtype2 9 64.6 (6.7)
subtype3 10 75.2 (9.2)
subtype4 13 68.8 (9.6)
subtype5 13 72.2 (9.8)

Figure S5.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

'METHLYATION CNMF' versus 'RACE'

P value = 0.955 (Fisher's exact test), Q value = 1

Table S8.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 44
subtype1 0 2 10
subtype2 0 1 8
subtype3 1 2 7
subtype4 1 3 9
subtype5 1 1 10

Figure S6.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'RACE'

Clustering Approach #3: 'RNAseq CNMF subtypes'

Table S9.  Description of clustering approach #3: 'RNAseq CNMF subtypes'

Cluster Labels 1 2 3 4
Number of samples 15 20 12 10
'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.99 (logrank test), Q value = 1

Table S10.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 56 34 0.3 - 102.4 (18.4)
subtype1 15 11 0.3 - 102.4 (18.8)
subtype2 19 10 3.7 - 59.4 (17.8)
subtype3 12 7 3.8 - 85.3 (16.5)
subtype4 10 6 2.7 - 93.4 (23.9)

Figure S7.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.205 (Kruskal-Wallis (anova)), Q value = 1

Table S11.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 57 69.7 (9.3)
subtype1 15 69.8 (8.9)
subtype2 20 71.9 (9.9)
subtype3 12 70.2 (10.2)
subtype4 10 64.7 (6.3)

Figure S8.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.658 (Fisher's exact test), Q value = 1

Table S12.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 44
subtype1 0 3 12
subtype2 2 2 16
subtype3 1 1 9
subtype4 0 3 7

Figure S9.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'RACE'

Clustering Approach #4: 'RNAseq cHierClus subtypes'

Table S13.  Description of clustering approach #4: 'RNAseq cHierClus subtypes'

Cluster Labels 1 2 3 4 5 6 7
Number of samples 11 4 8 6 8 6 14
'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0971 (logrank test), Q value = 1

Table S14.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 56 34 0.3 - 102.4 (18.4)
subtype1 11 7 0.3 - 102.4 (27.4)
subtype2 4 4 4.7 - 22.5 (9.0)
subtype3 8 4 11.9 - 59.4 (18.8)
subtype4 6 4 3.8 - 85.3 (14.8)
subtype5 8 4 2.7 - 69.6 (30.0)
subtype6 6 3 9.8 - 55.0 (26.7)
subtype7 13 8 3.7 - 93.4 (17.2)

Figure S10.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.234 (Kruskal-Wallis (anova)), Q value = 1

Table S15.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 57 69.7 (9.3)
subtype1 11 67.4 (8.0)
subtype2 4 71.2 (4.3)
subtype3 8 74.6 (8.1)
subtype4 6 66.8 (8.4)
subtype5 8 64.6 (6.8)
subtype6 6 73.0 (11.9)
subtype7 14 71.1 (11.2)

Figure S11.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.845 (Fisher's exact test), Q value = 1

Table S16.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 44
subtype1 0 3 8
subtype2 0 1 3
subtype3 1 2 5
subtype4 0 0 5
subtype5 0 1 7
subtype6 1 0 5
subtype7 1 2 11

Figure S12.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'RACE'

Clustering Approach #5: 'MIRSEQ CNMF'

Table S17.  Description of clustering approach #5: 'MIRSEQ CNMF'

Cluster Labels 1 2 3 4
Number of samples 11 22 4 19
'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.11 (logrank test), Q value = 1

Table S18.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 55 33 0.3 - 102.4 (18.1)
subtype1 11 8 4.7 - 102.4 (18.1)
subtype2 22 12 0.3 - 93.4 (25.0)
subtype3 4 1 2.7 - 69.6 (34.9)
subtype4 18 12 3.7 - 31.2 (15.5)

Figure S13.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 0.496 (Kruskal-Wallis (anova)), Q value = 1

Table S19.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 56 69.8 (9.3)
subtype1 11 69.4 (8.8)
subtype2 22 68.5 (9.3)
subtype3 4 67.5 (7.5)
subtype4 19 72.0 (10.2)

Figure S14.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.828 (Fisher's exact test), Q value = 1

Table S20.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 43
subtype1 0 2 9
subtype2 1 5 15
subtype3 0 0 4
subtype4 2 2 15

Figure S15.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'RACE'

Clustering Approach #6: 'MIRSEQ CHIERARCHICAL'

Table S21.  Description of clustering approach #6: 'MIRSEQ CHIERARCHICAL'

Cluster Labels 1 2 3 4
Number of samples 24 8 18 6
'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.935 (logrank test), Q value = 1

Table S22.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 55 33 0.3 - 102.4 (18.1)
subtype1 23 14 2.7 - 102.4 (17.2)
subtype2 8 6 0.3 - 74.4 (27.0)
subtype3 18 10 3.8 - 93.4 (23.0)
subtype4 6 3 13.4 - 25.7 (16.8)

Figure S16.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.209 (Kruskal-Wallis (anova)), Q value = 1

Table S23.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 56 69.8 (9.3)
subtype1 24 70.9 (9.8)
subtype2 8 67.4 (8.7)
subtype3 18 67.7 (9.1)
subtype4 6 74.8 (8.0)

Figure S17.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.81 (Fisher's exact test), Q value = 1

Table S24.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 43
subtype1 1 3 20
subtype2 0 2 6
subtype3 1 3 13
subtype4 1 1 4

Figure S18.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'RACE'

Clustering Approach #7: 'MIRseq Mature CNMF subtypes'

Table S25.  Description of clustering approach #7: 'MIRseq Mature CNMF subtypes'

Cluster Labels 1 2 3 4
Number of samples 11 23 5 17
'MIRseq Mature CNMF subtypes' versus 'Time to Death'

P value = 0.156 (logrank test), Q value = 1

Table S26.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 55 33 0.3 - 102.4 (18.1)
subtype1 11 9 3.8 - 102.4 (15.5)
subtype2 23 13 0.3 - 93.4 (26.6)
subtype3 5 1 2.7 - 69.6 (35.8)
subtype4 16 10 3.7 - 31.2 (16.5)

Figure S19.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.221 (Kruskal-Wallis (anova)), Q value = 1

Table S27.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 56 69.8 (9.3)
subtype1 11 68.3 (8.7)
subtype2 23 68.4 (9.1)
subtype3 5 68.2 (8.1)
subtype4 17 73.2 (10.2)

Figure S20.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.266 (Fisher's exact test), Q value = 1

Table S28.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 43
subtype1 0 0 11
subtype2 1 5 16
subtype3 0 2 3
subtype4 2 2 13

Figure S21.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'RACE'

Clustering Approach #8: 'MIRseq Mature cHierClus subtypes'

Table S29.  Description of clustering approach #8: 'MIRseq Mature cHierClus subtypes'

Cluster Labels 1 2 3 4 5 6
Number of samples 10 7 15 10 9 5
'MIRseq Mature cHierClus subtypes' versus 'Time to Death'

P value = 0.681 (logrank test), Q value = 1

Table S30.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 55 33 0.3 - 102.4 (18.1)
subtype1 10 7 2.7 - 102.4 (8.2)
subtype2 7 5 0.3 - 74.4 (27.4)
subtype3 14 8 3.7 - 69.6 (19.0)
subtype4 10 7 4.9 - 59.7 (25.0)
subtype5 9 3 9.8 - 93.4 (17.4)
subtype6 5 3 13.4 - 19.6 (15.8)

Figure S22.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.275 (Kruskal-Wallis (anova)), Q value = 1

Table S31.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 56 69.8 (9.3)
subtype1 10 68.1 (7.1)
subtype2 7 67.4 (9.4)
subtype3 15 72.2 (11.1)
subtype4 10 68.1 (9.4)
subtype5 9 67.7 (8.6)
subtype6 5 76.6 (7.5)

Figure S23.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.634 (Fisher's exact test), Q value = 1

Table S32.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'RACE'

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 3 9 43
subtype1 0 2 8
subtype2 0 1 6
subtype3 1 1 13
subtype4 1 2 7
subtype5 0 3 5
subtype6 1 0 4

Figure S24.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'RACE'

Methods & Data
Input
  • Cluster data file = UCS-TP.mergedcluster.txt

  • Clinical data file = UCS-TP.merged_data.txt

  • Number of patients = 57

  • Number of clustering approaches = 8

  • Number of selected clinical features = 3

  • Exclude small clusters that include fewer than K patients, K = 3

Clustering approaches
CNMF clustering

consensus non-negative matrix factorization clustering approach (Brunet et al. 2004)

Consensus hierarchical clustering

Resampling-based clustering method (Monti et al. 2003)

Survival analysis

For survival clinical features, the Kaplan-Meier survival curves of tumors with and without gene mutations were plotted and the statistical significance P values were estimated by logrank test (Bland and Altman 2004) using the 'survdiff' function in R

Fisher's exact test

For binary clinical features, two-tailed Fisher's exact tests (Fisher 1922) were used to estimate the P values using the 'fisher.test' function in R

Q value calculation

For multiple hypothesis correction, Q value is the False Discovery Rate (FDR) analogue of the P value (Benjamini and Hochberg 1995), defined as the minimum FDR at which the test may be called significant. We used the 'Benjamini and Hochberg' method of 'p.adjust' function in R to convert P values into Q values.

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.

References
[1] Brunet et al., Metagenes and molecular pattern discovery using matrix factorization, PNAS 101(12):4164-9 (2004)
[3] Bland and Altman, Statistics notes: The logrank test, BMJ 328(7447):1073 (2004)
[4] Fisher, R.A., On the interpretation of chi-square from contingency tables, and the calculation of P, Journal of the Royal Statistical Society 85(1):87-94 (1922)
[5] Benjamini and Hochberg, Controlling the false discovery rate: a practical and powerful approach to multiple testing, Journal of the Royal Statistical Society Series B 59:289-300 (1995)