Correlation between aggregated molecular cancer subtypes and selected clinical features
Uveal Melanoma (Primary solid tumor)
28 January 2016  |  analyses__2016_01_28
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 (2016): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C19Z94FD
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 7 clinical features across 80 patients, 10 significant findings detected with P value < 0.05 and Q value < 0.25.

  • 6 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'Time to Death'.

  • 4 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time to Death'.

  • CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that correlate to 'Time to Death'.

  • Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time to Death' and 'PATHOLOGIC_STAGE'.

  • 6 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'Time to Death'.

  • 4 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'Time to Death'.

  • 7 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'Time to Death' and 'YEARS_TO_BIRTH'.

  • 4 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'Time to Death'.

Results
Overview of the results

Table 1.  Get Full Table Overview of the association between subtypes identified by 8 different clustering approaches and 7 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 10 significant findings detected.

Clinical
Features
Time
to
Death
YEARS
TO
BIRTH
PATHOLOGIC
STAGE
PATHOLOGY
T
STAGE
PATHOLOGY
M
STAGE
GENDER RADIATION
THERAPY
Statistical Tests logrank test Kruskal-Wallis (anova) Fisher's exact test Fisher's exact test Fisher's exact test Fisher's exact test Fisher's exact test
Copy Number Ratio CNMF subtypes 7.12e-05
(0.000498)
0.0805
(0.237)
0.225
(0.34)
0.393
(0.5)
0.0902
(0.237)
0.983
(1.00)
0.121
(0.238)
METHLYATION CNMF 2.71e-07
(7.59e-06)
0.0815
(0.237)
0.352
(0.458)
0.603
(0.719)
0.181
(0.315)
0.172
(0.311)
0.773
(0.848)
RNAseq CNMF subtypes 3.45e-06
(3.22e-05)
0.331
(0.45)
0.0545
(0.237)
0.0548
(0.237)
0.0701
(0.237)
0.658
(0.768)
0.894
(0.945)
RNAseq cHierClus subtypes 2.56e-08
(1.43e-06)
0.0852
(0.237)
0.0407
(0.228)
0.116
(0.238)
0.0651
(0.237)
0.878
(0.945)
1
(1.00)
MIRSEQ CNMF 2.18e-06
(2.67e-05)
0.111
(0.238)
0.31
(0.445)
0.127
(0.238)
0.28
(0.412)
0.715
(0.817)
0.753
(0.844)
MIRSEQ CHIERARCHICAL 2.38e-06
(2.67e-05)
0.125
(0.238)
0.192
(0.315)
0.0974
(0.237)
0.12
(0.238)
0.199
(0.318)
1
(1.00)
MIRseq Mature CNMF subtypes 5.15e-07
(9.62e-06)
0.00434
(0.027)
0.191
(0.315)
0.11
(0.238)
0.0968
(0.237)
0.0575
(0.237)
0.338
(0.45)
MIRseq Mature cHierClus subtypes 4.33e-06
(3.46e-05)
0.0762
(0.237)
0.216
(0.336)
0.335
(0.45)
0.0779
(0.237)
0.518
(0.63)
0.425
(0.529)
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 4 5 6
Number of samples 22 7 10 31 6 4
'Copy Number Ratio CNMF subtypes' versus 'Time to Death'

P value = 7.12e-05 (logrank test), Q value = 5e-04

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

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 22 7 0.1 - 82.2 (26.2)
subtype2 7 4 2.7 - 24.0 (13.6)
subtype3 10 6 13.1 - 45.9 (19.8)
subtype4 31 2 0.2 - 85.5 (33.8)
subtype5 6 2 1.4 - 36.6 (30.5)
subtype6 4 2 1.6 - 43.2 (23.0)

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 'YEARS_TO_BIRTH'

P value = 0.0805 (Kruskal-Wallis (anova)), Q value = 0.24

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

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 22 60.5 (11.0)
subtype2 7 72.6 (12.2)
subtype3 10 65.5 (15.3)
subtype4 31 57.5 (14.7)
subtype5 6 61.3 (14.7)
subtype6 4 71.8 (10.2)

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.225 (Fisher's exact test), Q value = 0.34

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

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 1 7 9 3 0 2
subtype2 0 1 2 1 1 2
subtype3 2 4 2 2 0 0
subtype4 7 13 8 3 0 0
subtype5 1 2 3 0 0 0
subtype6 1 0 1 1 0 0

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.393 (Fisher's exact test), Q value = 0.5

Table S5.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 2 9 11
subtype2 1 0 6
subtype3 2 4 4
subtype4 7 14 10
subtype5 1 3 2
subtype6 1 2 1

Figure S4.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0902 (Fisher's exact test), Q value = 0.24

Table S6.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 12 2
subtype2 4 2
subtype3 6 0
subtype4 22 0
subtype5 4 0
subtype6 3 0

Figure S5.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.983 (Fisher's exact test), Q value = 1

Table S7.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 11 11
subtype2 3 4
subtype3 4 6
subtype4 13 18
subtype5 2 4
subtype6 2 2

Figure S6.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'GENDER'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.121 (Fisher's exact test), Q value = 0.24

Table S8.  Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 21 0
subtype2 6 1
subtype3 10 0
subtype4 30 1
subtype5 6 0
subtype6 3 1

Figure S7.  Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #2: 'METHLYATION CNMF'

Table S9.  Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels 1 2 3 4
Number of samples 28 14 14 24
'METHLYATION CNMF' versus 'Time to Death'

P value = 2.71e-07 (logrank test), Q value = 7.6e-06

Table S10.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 28 17 0.1 - 45.3 (19.3)
subtype2 14 1 0.2 - 85.5 (34.2)
subtype3 14 3 1.3 - 82.2 (24.8)
subtype4 24 2 0.2 - 46.8 (27.1)

Figure S8.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.0815 (Kruskal-Wallis (anova)), Q value = 0.24

Table S11.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 28 66.9 (12.1)
subtype2 14 60.2 (12.6)
subtype3 14 61.9 (13.4)
subtype4 24 56.2 (15.5)

Figure S9.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.352 (Fisher's exact test), Q value = 0.46

Table S12.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 1 10 8 3 1 4
subtype2 2 4 6 2 0 0
subtype3 3 3 6 2 0 0
subtype4 6 10 5 3 0 0

Figure S10.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.603 (Fisher's exact test), Q value = 0.72

Table S13.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 2 11 15
subtype2 3 5 6
subtype3 3 6 5
subtype4 6 10 8

Figure S11.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.181 (Fisher's exact test), Q value = 0.32

Table S14.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 18 4
subtype2 8 0
subtype3 9 0
subtype4 16 0

Figure S12.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.172 (Fisher's exact test), Q value = 0.31

Table S15.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 11 17
subtype2 5 9
subtype3 10 4
subtype4 9 15

Figure S13.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'GENDER'

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.773 (Fisher's exact test), Q value = 0.85

Table S16.  Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 25 2
subtype2 14 0
subtype3 14 0
subtype4 23 1

Figure S14.  Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #3: 'RNAseq CNMF subtypes'

Table S17.  Description of clustering approach #3: 'RNAseq CNMF subtypes'

Cluster Labels 1 2 3 4
Number of samples 18 15 32 15
'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 3.45e-06 (logrank test), Q value = 3.2e-05

Table S18.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 18 9 1.3 - 49.7 (14.9)
subtype2 15 5 0.4 - 61.2 (27.5)
subtype3 32 1 0.2 - 85.5 (36.4)
subtype4 15 8 0.1 - 45.3 (22.0)

Figure S15.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.331 (Kruskal-Wallis (anova)), Q value = 0.45

Table S19.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 18 67.2 (14.3)
subtype2 15 61.0 (13.1)
subtype3 32 59.0 (14.4)
subtype4 15 61.4 (12.8)

Figure S16.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0545 (Fisher's exact test), Q value = 0.24

Table S20.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 2 2 7 3 1 3
subtype2 0 6 5 3 0 0
subtype3 9 12 8 3 0 0
subtype4 1 7 5 1 0 1

Figure S17.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0548 (Fisher's exact test), Q value = 0.24

Table S21.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 2 3 13
subtype2 2 8 5
subtype3 9 14 9
subtype4 1 7 7

Figure S18.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0701 (Fisher's exact test), Q value = 0.24

Table S22.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 12 3
subtype2 10 0
subtype3 20 0
subtype4 9 1

Figure S19.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.658 (Fisher's exact test), Q value = 0.77

Table S23.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 9 9
subtype2 8 7
subtype3 13 19
subtype4 5 10

Figure S20.  Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.894 (Fisher's exact test), Q value = 0.95

Table S24.  Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 16 1
subtype2 14 1
subtype3 31 1
subtype4 15 0

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

Clustering Approach #4: 'RNAseq cHierClus subtypes'

Table S25.  Description of clustering approach #4: 'RNAseq cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 29 33 18
'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 2.56e-08 (logrank test), Q value = 1.4e-06

Table S26.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 29 17 0.1 - 49.7 (15.4)
subtype2 33 1 0.2 - 85.5 (33.8)
subtype3 18 5 0.4 - 61.2 (29.3)

Figure S22.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0852 (Kruskal-Wallis (anova)), Q value = 0.24

Table S27.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 29 66.4 (12.5)
subtype2 33 57.8 (14.5)
subtype3 18 60.9 (13.4)

Figure S23.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0407 (Fisher's exact test), Q value = 0.23

Table S28.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 3 9 9 3 1 4
subtype2 9 13 8 3 0 0
subtype3 0 5 8 4 0 0

Figure S24.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.116 (Fisher's exact test), Q value = 0.24

Table S29.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 4 9 16
subtype2 9 15 9
subtype3 1 8 9

Figure S25.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0651 (Fisher's exact test), Q value = 0.24

Table S30.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 18 4
subtype2 22 0
subtype3 11 0

Figure S26.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.878 (Fisher's exact test), Q value = 0.95

Table S31.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 12 17
subtype2 14 19
subtype3 9 9

Figure S27.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

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

Table S32.  Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 27 1
subtype2 32 1
subtype3 17 1

Figure S28.  Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #5: 'MIRSEQ CNMF'

Table S33.  Description of clustering approach #5: 'MIRSEQ CNMF'

Cluster Labels 1 2 3 4 5 6
Number of samples 8 8 11 25 23 5
'MIRSEQ CNMF' versus 'Time to Death'

P value = 2.18e-06 (logrank test), Q value = 2.7e-05

Table S34.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 8 3 1.4 - 45.9 (33.3)
subtype2 8 3 7.9 - 61.2 (33.3)
subtype3 11 0 0.2 - 85.5 (31.8)
subtype4 25 1 0.2 - 82.2 (27.0)
subtype5 23 14 0.1 - 45.3 (18.9)
subtype6 5 2 1.3 - 49.7 (20.9)

Figure S29.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.111 (Kruskal-Wallis (anova)), Q value = 0.24

Table S35.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 8 56.9 (16.2)
subtype2 8 64.5 (11.6)
subtype3 11 57.4 (11.3)
subtype4 25 57.6 (15.9)
subtype5 23 66.9 (11.5)
subtype6 5 70.2 (10.7)

Figure S30.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.31 (Fisher's exact test), Q value = 0.44

Table S36.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 0 4 2 1 0 1
subtype2 0 2 3 2 0 0
subtype3 0 4 6 1 0 0
subtype4 9 8 5 3 0 0
subtype5 2 7 8 2 1 3
subtype6 1 2 1 1 0 0

Figure S31.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.127 (Fisher's exact test), Q value = 0.24

Table S37.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 1 4 3
subtype2 1 4 3
subtype3 0 4 7
subtype4 9 10 6
subtype5 2 7 14
subtype6 1 3 1

Figure S32.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.28 (Fisher's exact test), Q value = 0.41

Table S38.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 4 1
subtype2 3 0
subtype3 9 0
subtype4 17 0
subtype5 14 3
subtype6 4 0

Figure S33.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.715 (Fisher's exact test), Q value = 0.82

Table S39.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 5 3
subtype2 3 5
subtype3 4 7
subtype4 12 13
subtype5 8 15
subtype6 3 2

Figure S34.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'GENDER'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY'

P value = 0.753 (Fisher's exact test), Q value = 0.84

Table S40.  Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 7 0
subtype2 7 1
subtype3 11 0
subtype4 24 1
subtype5 22 1
subtype6 5 0

Figure S35.  Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #6: 'MIRSEQ CHIERARCHICAL'

Table S41.  Description of clustering approach #6: 'MIRSEQ CHIERARCHICAL'

Cluster Labels 1 2 3 4
Number of samples 17 13 28 22
'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 2.38e-06 (logrank test), Q value = 2.7e-05

Table S42.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 80 23 0.1 - 85.5 (25.8)
subtype1 17 7 1.4 - 61.2 (28.7)
subtype2 13 1 0.2 - 85.5 (37.8)
subtype3 28 2 0.2 - 82.2 (26.8)
subtype4 22 13 0.1 - 49.7 (14.3)

Figure S36.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.125 (Kruskal-Wallis (anova)), Q value = 0.24

Table S43.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 80 61.6 (13.9)
subtype1 17 60.9 (14.9)
subtype2 13 59.2 (12.2)
subtype3 28 58.4 (15.7)
subtype4 22 67.8 (10.2)

Figure S37.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGIC_STAGE'

P value = 0.192 (Fisher's exact test), Q value = 0.32

Table S44.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 12 27 25 10 1 4
subtype1 1 7 5 1 1 1
subtype2 0 6 5 2 0 0
subtype3 9 8 7 4 0 0
subtype4 2 6 8 3 0 3

Figure S38.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.0974 (Fisher's exact test), Q value = 0.24

Table S45.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 14 32 34
subtype1 3 9 5
subtype2 0 6 7
subtype3 9 10 9
subtype4 2 7 13

Figure S39.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.12 (Fisher's exact test), Q value = 0.24

Table S46.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 51 4
subtype1 10 1
subtype2 9 0
subtype3 19 0
subtype4 13 3

Figure S40.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.199 (Fisher's exact test), Q value = 0.32

Table S47.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 35 45
subtype1 10 7
subtype2 3 10
subtype3 14 14
subtype4 8 14

Figure S41.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY'

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

Table S48.  Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 76 3
subtype1 15 1
subtype2 13 0
subtype3 27 1
subtype4 21 1

Figure S42.  Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #7: 'MIRseq Mature CNMF subtypes'

Table S49.  Description of clustering approach #7: 'MIRseq Mature CNMF subtypes'

Cluster Labels 1 2 3 4 5 6 7
Number of samples 20 7 17 9 8 7 6
'MIRseq Mature CNMF subtypes' versus 'Time to Death'

P value = 5.15e-07 (logrank test), Q value = 9.6e-06

Table S50.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 74 23 0.1 - 85.5 (25.0)
subtype1 20 14 0.1 - 41.7 (16.9)
subtype2 7 2 0.4 - 61.2 (36.6)
subtype3 17 1 14.0 - 85.5 (38.7)
subtype4 9 1 2.1 - 41.0 (35.4)
subtype5 8 1 1.6 - 56.0 (24.3)
subtype6 7 1 0.2 - 27.6 (19.7)
subtype7 6 3 1.3 - 46.8 (35.8)

Figure S43.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00434 (Kruskal-Wallis (anova)), Q value = 0.027

Table S51.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 74 61.9 (14.1)
subtype1 20 66.7 (13.1)
subtype2 7 58.7 (17.0)
subtype3 17 52.3 (15.5)
subtype4 9 54.4 (9.1)
subtype5 8 69.4 (10.2)
subtype6 7 71.0 (7.4)
subtype7 6 67.8 (6.5)

Figure S44.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.191 (Fisher's exact test), Q value = 0.32

Table S52.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 10 25 24 9 1 4
subtype1 1 6 8 1 1 3
subtype2 0 4 1 1 0 0
subtype3 2 7 6 2 0 0
subtype4 0 1 5 3 0 0
subtype5 3 2 3 0 0 0
subtype6 2 3 1 1 0 0
subtype7 2 2 0 1 0 1

Figure S45.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.11 (Fisher's exact test), Q value = 0.24

Table S53.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 12 30 32
subtype1 1 7 12
subtype2 2 4 1
subtype3 2 8 7
subtype4 0 2 7
subtype5 3 4 1
subtype6 2 3 2
subtype7 2 2 2

Figure S46.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0968 (Fisher's exact test), Q value = 0.24

Table S54.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 48 4
subtype1 14 3
subtype2 4 0
subtype3 14 0
subtype4 5 0
subtype5 7 0
subtype6 4 0
subtype7 0 1

Figure S47.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.0575 (Fisher's exact test), Q value = 0.24

Table S55.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 32 42
subtype1 9 11
subtype2 3 4
subtype3 4 13
subtype4 2 7
subtype5 7 1
subtype6 3 4
subtype7 4 2

Figure S48.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.338 (Fisher's exact test), Q value = 0.45

Table S56.  Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 70 3
subtype1 18 1
subtype2 6 1
subtype3 17 0
subtype4 9 0
subtype5 8 0
subtype6 6 1
subtype7 6 0

Figure S49.  Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

Clustering Approach #8: 'MIRseq Mature cHierClus subtypes'

Table S57.  Description of clustering approach #8: 'MIRseq Mature cHierClus subtypes'

Cluster Labels 1 2 3 4
Number of samples 18 28 16 12
'MIRseq Mature cHierClus subtypes' versus 'Time to Death'

P value = 4.33e-06 (logrank test), Q value = 3.5e-05

Table S58.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 74 23 0.1 - 85.5 (25.0)
subtype1 18 9 1.4 - 61.2 (23.7)
subtype2 28 2 0.2 - 85.5 (30.7)
subtype3 16 10 0.1 - 49.7 (14.3)
subtype4 12 2 0.4 - 45.9 (29.6)

Figure S50.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0762 (Kruskal-Wallis (anova)), Q value = 0.24

Table S59.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

nPatients Mean (Std.Dev)
ALL 74 61.9 (14.1)
subtype1 18 64.1 (12.5)
subtype2 28 59.2 (15.4)
subtype3 16 68.8 (11.6)
subtype4 12 55.9 (13.7)

Figure S51.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.216 (Fisher's exact test), Q value = 0.34

Table S60.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV
ALL 10 25 24 9 1 4
subtype1 1 3 7 3 1 2
subtype2 7 11 7 3 0 0
subtype3 2 7 4 1 0 2
subtype4 0 4 6 2 0 0

Figure S52.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.335 (Fisher's exact test), Q value = 0.45

Table S61.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients T2 T3 T4
ALL 12 30 32
subtype1 2 6 10
subtype2 7 13 8
subtype3 2 8 6
subtype4 1 3 8

Figure S53.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0779 (Fisher's exact test), Q value = 0.24

Table S62.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

nPatients 0 1
ALL 48 4
subtype1 10 2
subtype2 21 0
subtype3 10 2
subtype4 7 0

Figure S54.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_M_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.518 (Fisher's exact test), Q value = 0.63

Table S63.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'GENDER'

nPatients FEMALE MALE
ALL 32 42
subtype1 6 12
subtype2 11 17
subtype3 8 8
subtype4 7 5

Figure S55.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.425 (Fisher's exact test), Q value = 0.53

Table S64.  Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

nPatients NO YES
ALL 70 3
subtype1 15 2
subtype2 27 1
subtype3 16 0
subtype4 12 0

Figure S56.  Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'RADIATION_THERAPY'

Methods & Data
Input
  • Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/UVM-TP/22542449/UVM-TP.mergedcluster.txt

  • Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/UVM-TP/22507229/UVM-TP.merged_data.txt

  • Number of patients = 80

  • Number of clustering approaches = 8

  • Number of selected clinical features = 7

  • 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)