Correlation between aggregated molecular cancer subtypes and selected clinical features

Glioblastoma Multiforme (Primary solid tumor)

21 August 2015 | analyses__2015_08_21

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 (2015): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1BP020W

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 10 different clustering approaches and 8 clinical features across 593 patients, 12 significant findings detected with P value < 0.05 and Q value < 0.25.

CNMF clustering analysis on array-based mRNA expression data identified 4 subtypes that correlate to 'HISTOLOGICAL_TYPE'.
Consensus hierarchical clustering analysis on array-based mRNA expression data identified 4 subtypes that correlate to 'Time to Death', 'YEARS_TO_BIRTH', and 'HISTOLOGICAL_TYPE'.
CNMF clustering analysis on array-based miR expression data identified 4 subtypes that correlate to 'Time to Death'.
Consensus hierarchical clustering analysis on array-based miR expression data identified 3 subtypes that do not correlate to any clinical features.
3 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'YEARS_TO_BIRTH' and 'RADIATION_THERAPY'.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time to Death', 'YEARS_TO_BIRTH', and 'KARNOFSKY_PERFORMANCE_SCORE'.
CNMF clustering analysis on RPPA data identified 5 subtypes that do not correlate to any clinical features.
Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that correlate to 'Time to Death'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that do not correlate to any clinical features.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'KARNOFSKY_PERFORMANCE_SCORE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 10 different clustering approaches and 8 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 12 significant findings detected.


Clinical Features	Time to Death	YEARS TO BIRTH	GENDER	RADIATION THERAPY	KARNOFSKY PERFORMANCE SCORE	HISTOLOGICAL TYPE	RACE	ETHNICITY
Statistical Tests	logrank test	Kruskal-Wallis (anova)	Fisher's exact test	Fisher's exact test	Kruskal-Wallis (anova)	Fisher's exact test	Fisher's exact test	Fisher's exact test
mRNA CNMF subtypes	0.0522 (0.278)	0.0788 (0.3)	0.4 (0.745)	0.517 (0.811)	0.474 (0.79)	0.00474 (0.0542)	0.834 (0.93)	0.123 (0.377)
mRNA cHierClus subtypes	0.00017 (0.00452)	0.00252 (0.0403)	0.0926 (0.335)	0.224 (0.528)	0.515 (0.811)	0.0302 (0.22)	0.586 (0.876)	0.682 (0.902)
miR CNMF subtypes	0.00335 (0.0447)	0.38 (0.724)	0.651 (0.902)	0.352 (0.705)	0.861 (0.93)	0.362 (0.707)	0.205 (0.513)	0.855 (0.93)
miR cHierClus subtypes	0.194 (0.5)	0.127 (0.377)	0.151 (0.417)	0.463 (0.788)	0.278 (0.591)	0.281 (0.591)	0.192 (0.5)	0.675 (0.902)
Copy Number Ratio CNMF subtypes	0.0435 (0.255)	1.21e-05 (0.00097)	0.49 (0.8)	0.00551 (0.0551)	0.103 (0.345)	0.224 (0.528)	0.12 (0.377)	0.759 (0.906)
METHLYATION CNMF	0.000425 (0.0085)	8.57e-05 (0.00343)	0.601 (0.876)	0.238 (0.528)	0.0301 (0.22)	0.302 (0.619)	0.992 (0.992)	0.893 (0.933)
RPPA CNMF subtypes	0.688 (0.902)	0.856 (0.93)	0.0616 (0.3)	0.767 (0.906)	0.988 (0.992)	0.896 (0.933)	0.987 (0.992)	0.668 (0.902)
RPPA cHierClus subtypes	0.0238 (0.211)	0.0447 (0.255)	0.236 (0.528)	0.647 (0.902)	0.068 (0.3)	0.732 (0.906)	0.42 (0.747)	0.781 (0.906)
RNAseq CNMF subtypes	0.45 (0.782)	0.0964 (0.335)	0.0701 (0.3)	0.706 (0.906)	0.41 (0.745)	0.791 (0.906)	0.792 (0.906)	0.792 (0.906)
RNAseq cHierClus subtypes	0.0741 (0.3)	0.0784 (0.3)	0.145 (0.413)	0.898 (0.933)	0.0366 (0.244)	0.602 (0.876)	0.577 (0.876)	0.738 (0.906)

Clustering Approach #1: 'mRNA CNMF subtypes'

Table S1. Description of clustering approach #1: 'mRNA CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	154	97	156	118

'mRNA CNMF subtypes' versus 'Time to Death'

P value = 0.0522 (logrank test), Q value = 0.28

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	525	447	0.1 - 127.6 (12.2)
subtype1	154	138	0.1 - 127.6 (10.6)
subtype2	97	81	0.2 - 115.9 (14.5)
subtype3	156	128	0.1 - 94.8 (13.8)
subtype4	118	100	0.2 - 91.8 (12.6)

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

'mRNA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0788 (Kruskal-Wallis (anova)), Q value = 0.3

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

	nPatients	Mean (Std.Dev)
ALL	525	57.7 (14.6)
subtype1	154	58.4 (12.4)
subtype2	97	54.2 (17.4)
subtype3	156	60.0 (13.3)
subtype4	118	56.5 (15.7)

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

'mRNA CNMF subtypes' versus 'GENDER'

P value = 0.4 (Fisher's exact test), Q value = 0.74

Table S4. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	205	320
subtype1	59	95
subtype2	41	56
subtype3	66	90
subtype4	39	79

Figure S3. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'mRNA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.517 (Fisher's exact test), Q value = 0.81

Table S5. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	70	435
subtype1	21	127
subtype2	14	79
subtype3	16	134
subtype4	19	95

Figure S4. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'mRNA CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.474 (Kruskal-Wallis (anova)), Q value = 0.79

Table S6. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	396	77.1 (14.6)
subtype1	115	77.6 (15.8)
subtype2	76	76.6 (11.0)
subtype3	120	76.1 (15.4)
subtype4	85	78.6 (14.7)

Figure S5. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'mRNA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.00474 (Fisher's exact test), Q value = 0.054

Table S7. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	6	20	499
subtype1	1	8	145
subtype2	0	7	90
subtype3	0	4	152
subtype4	5	1	112

Figure S6. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'mRNA CNMF subtypes' versus 'RACE'

P value = 0.834 (Fisher's exact test), Q value = 0.93

Table S8. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	13	31	462
subtype1	4	12	135
subtype2	2	7	84
subtype3	5	7	139
subtype4	2	5	104

Figure S7. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #7: 'RACE'

'mRNA CNMF subtypes' versus 'ETHNICITY'

P value = 0.123 (Fisher's exact test), Q value = 0.38

Table S9. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	12	438
subtype1	3	131
subtype2	5	68
subtype3	3	135
subtype4	1	104

Figure S8. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #2: 'mRNA cHierClus subtypes'

Table S10. Description of clustering approach #2: 'mRNA cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	153	107	103	162

'mRNA cHierClus subtypes' versus 'Time to Death'

P value = 0.00017 (logrank test), Q value = 0.0045

Table S11. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	525	447	0.1 - 127.6 (12.2)
subtype1	153	138	0.1 - 91.0 (12.2)
subtype2	107	80	0.2 - 115.9 (14.7)
subtype3	103	87	0.1 - 94.8 (13.8)
subtype4	162	142	0.1 - 127.6 (10.4)

Figure S9. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'mRNA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00252 (Kruskal-Wallis (anova)), Q value = 0.04

Table S12. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	525	57.7 (14.6)
subtype1	153	56.9 (13.6)
subtype2	107	52.9 (17.9)
subtype3	103	60.6 (12.1)
subtype4	162	59.8 (13.7)

Figure S10. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'mRNA cHierClus subtypes' versus 'GENDER'

P value = 0.0926 (Fisher's exact test), Q value = 0.34

Table S13. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	205	320
subtype1	51	102
subtype2	44	63
subtype3	50	53
subtype4	60	102

Figure S11. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'mRNA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.224 (Fisher's exact test), Q value = 0.53

Table S14. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	70	435
subtype1	17	131
subtype2	13	91
subtype3	11	88
subtype4	29	125

Figure S12. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'mRNA cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.515 (Kruskal-Wallis (anova)), Q value = 0.81

Table S15. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	396	77.1 (14.6)
subtype1	118	78.1 (15.4)
subtype2	84	77.6 (11.6)
subtype3	81	75.2 (15.6)
subtype4	113	77.3 (15.1)

Figure S13. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'mRNA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.0302 (Fisher's exact test), Q value = 0.22

Table S16. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	6	20	499
subtype1	0	11	142
subtype2	1	3	103
subtype3	0	3	100
subtype4	5	3	154

Figure S14. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'mRNA cHierClus subtypes' versus 'RACE'

P value = 0.586 (Fisher's exact test), Q value = 0.88

Table S17. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	13	31	462
subtype1	4	13	130
subtype2	4	6	95
subtype3	3	4	93
subtype4	2	8	144

Figure S15. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'RACE'

'mRNA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.682 (Fisher's exact test), Q value = 0.9

Table S18. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	12	438
subtype1	5	124
subtype2	1	91
subtype3	2	90
subtype4	4	133

Figure S16. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #3: 'miR CNMF subtypes'

Table S19. Description of clustering approach #3: 'miR CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	165	182	90	124

'miR CNMF subtypes' versus 'Time to Death'

P value = 0.00335 (logrank test), Q value = 0.045

Table S20. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	561	466	0.1 - 127.6 (12.2)
subtype1	165	141	0.1 - 91.0 (11.3)
subtype2	182	147	0.1 - 127.6 (13.7)
subtype3	90	74	0.4 - 65.3 (9.3)
subtype4	124	104	0.1 - 120.6 (12.8)

Figure S17. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'miR CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.38 (Kruskal-Wallis (anova)), Q value = 0.72

Table S21. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	561	57.9 (14.3)
subtype1	165	59.8 (11.6)
subtype2	182	55.7 (16.6)
subtype3	90	58.5 (14.6)
subtype4	124	58.1 (13.6)

Figure S18. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'miR CNMF subtypes' versus 'GENDER'

P value = 0.651 (Fisher's exact test), Q value = 0.9

Table S22. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	218	343
subtype1	63	102
subtype2	76	106
subtype3	36	54
subtype4	43	81

Figure S19. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'miR CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.352 (Fisher's exact test), Q value = 0.7

Table S23. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	73	465
subtype1	19	141
subtype2	21	153
subtype3	17	70
subtype4	16	101

Figure S20. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'miR CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.861 (Kruskal-Wallis (anova)), Q value = 0.93

Table S24. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	428	77.4 (14.7)
subtype1	129	76.1 (15.5)
subtype2	132	77.4 (14.8)
subtype3	74	78.2 (13.5)
subtype4	93	78.3 (14.6)

Figure S21. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'miR CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.362 (Fisher's exact test), Q value = 0.71

Table S25. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	9	20	532
subtype1	2	4	159
subtype2	1	8	173
subtype3	3	5	82
subtype4	3	3	118

Figure S22. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'miR CNMF subtypes' versus 'RACE'

P value = 0.205 (Fisher's exact test), Q value = 0.51

Table S26. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	13	30	493
subtype1	3	13	140
subtype2	7	9	157
subtype3	1	1	85
subtype4	2	7	111

Figure S23. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #7: 'RACE'

'miR CNMF subtypes' versus 'ETHNICITY'

P value = 0.855 (Fisher's exact test), Q value = 0.93

Table S27. Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	12	459
subtype1	5	134
subtype2	4	153
subtype3	1	71
subtype4	2	101

Figure S24. Get High-res Image Clustering Approach #3: 'miR CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #4: 'miR cHierClus subtypes'

Table S28. Description of clustering approach #4: 'miR cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	302	129	130

'miR cHierClus subtypes' versus 'Time to Death'

P value = 0.194 (logrank test), Q value = 0.5

Table S29. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	561	466	0.1 - 127.6 (12.2)
subtype1	302	252	0.1 - 120.6 (12.6)
subtype2	129	107	0.1 - 92.7 (10.8)
subtype3	130	107	0.1 - 127.6 (11.8)

Figure S25. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'miR cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.127 (Kruskal-Wallis (anova)), Q value = 0.38

Table S30. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	561	57.9 (14.3)
subtype1	302	56.5 (15.9)
subtype2	129	59.0 (11.5)
subtype3	130	60.0 (12.5)

Figure S26. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'miR cHierClus subtypes' versus 'GENDER'

P value = 0.151 (Fisher's exact test), Q value = 0.42

Table S31. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	218	343
subtype1	110	192
subtype2	48	81
subtype3	60	70

Figure S27. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'miR cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.463 (Fisher's exact test), Q value = 0.79

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

nPatients	NO	YES
ALL	73	465
subtype1	44	245
subtype2	13	109
subtype3	16	111

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

'miR cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.278 (Kruskal-Wallis (anova)), Q value = 0.59

Table S33. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	428	77.4 (14.7)
subtype1	227	78.5 (14.0)
subtype2	101	76.5 (15.6)
subtype3	100	75.5 (15.4)

Figure S29. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'miR cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.281 (Fisher's exact test), Q value = 0.59

Table S34. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	9	20	532
subtype1	6	13	283
subtype2	3	2	124
subtype3	0	5	125

Figure S30. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'miR cHierClus subtypes' versus 'RACE'

P value = 0.192 (Fisher's exact test), Q value = 0.5

Table S35. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	13	30	493
subtype1	7	10	272
subtype2	3	11	113
subtype3	3	9	108

Figure S31. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #7: 'RACE'

'miR cHierClus subtypes' versus 'ETHNICITY'

P value = 0.675 (Fisher's exact test), Q value = 0.9

Table S36. Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	12	459
subtype1	6	249
subtype2	2	109
subtype3	4	101

Figure S32. Get High-res Image Clustering Approach #4: 'miR cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #5: 'Copy Number Ratio CNMF subtypes'

Table S37. Description of clustering approach #5: 'Copy Number Ratio CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	161	217	193

'Copy Number Ratio CNMF subtypes' versus 'Time to Death'

P value = 0.0435 (logrank test), Q value = 0.26

Table S38. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	570	470	0.1 - 127.6 (12.2)
subtype1	161	140	0.1 - 127.6 (10.6)
subtype2	217	179	0.1 - 120.6 (12.6)
subtype3	192	151	0.2 - 115.9 (12.6)

Figure S33. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'Copy Number Ratio CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 1.21e-05 (Kruskal-Wallis (anova)), Q value = 0.00097

Table S39. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	571	57.7 (14.5)
subtype1	161	60.5 (13.5)
subtype2	217	60.1 (10.7)
subtype3	193	52.9 (17.5)

Figure S34. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.49 (Fisher's exact test), Q value = 0.8

Table S40. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	224	347
subtype1	69	92
subtype2	80	137
subtype3	75	118

Figure S35. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.00551 (Fisher's exact test), Q value = 0.055

Table S41. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	74	474
subtype1	33	122
subtype2	22	185
subtype3	19	167

Figure S36. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'Copy Number Ratio CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.103 (Kruskal-Wallis (anova)), Q value = 0.34

Table S42. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	424	77.4 (14.9)
subtype1	115	75.7 (14.6)
subtype2	166	77.3 (15.6)
subtype3	143	79.0 (14.3)

Figure S37. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.224 (Fisher's exact test), Q value = 0.53

Table S43. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	28	18	525
subtype1	5	4	152
subtype2	9	5	203
subtype3	14	9	170

Figure S38. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.12 (Fisher's exact test), Q value = 0.38

Table S44. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	13	50	485
subtype1	1	15	139
subtype2	3	17	188
subtype3	9	18	158

Figure S39. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RACE'

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

P value = 0.759 (Fisher's exact test), Q value = 0.91

Table S45. Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	12	473
subtype1	2	134
subtype2	5	176
subtype3	5	163

Figure S40. Get High-res Image Clustering Approach #5: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #6: 'METHLYATION CNMF'

Table S46. Description of clustering approach #6: 'METHLYATION CNMF'

Cluster Labels	1	2	3
Number of samples	92	110	81

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.000425 (logrank test), Q value = 0.0085

Table S47. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	283	227	0.1 - 127.6 (12.2)
subtype1	92	73	0.1 - 92.7 (11.8)
subtype2	110	94	0.1 - 77.7 (10.5)
subtype3	81	60	0.2 - 127.6 (14.3)

Figure S41. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 8.57e-05 (Kruskal-Wallis (anova)), Q value = 0.0034

Table S48. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	283	57.8 (15.0)
subtype1	92	59.0 (12.9)
subtype2	110	61.8 (12.6)
subtype3	81	51.1 (17.8)

Figure S42. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.601 (Fisher's exact test), Q value = 0.88

Table S49. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	115	168
subtype1	41	51
subtype2	44	66
subtype3	30	51

Figure S43. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.238 (Fisher's exact test), Q value = 0.53

Table S50. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	39	234
subtype1	10	79
subtype2	20	85
subtype3	9	70

Figure S44. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #4: 'RADIATION_THERAPY'

'METHLYATION CNMF' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0301 (Kruskal-Wallis (anova)), Q value = 0.22

Table S51. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	216	75.4 (15.3)
subtype1	71	77.0 (17.7)
subtype2	81	72.6 (14.6)
subtype3	64	77.2 (13.0)

Figure S45. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.302 (Fisher's exact test), Q value = 0.62

Table S52. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	3	19	261
subtype1	3	6	83
subtype2	0	7	103
subtype3	0	6	75

Figure S46. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'RACE'

P value = 0.992 (Fisher's exact test), Q value = 0.99

Table S53. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	10	16	254
subtype1	3	6	82
subtype2	4	6	98
subtype3	3	4	74

Figure S47. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #7: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.893 (Fisher's exact test), Q value = 0.93

Table S54. Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	7	232
subtype1	3	80
subtype2	2	88
subtype3	2	64

Figure S48. Get High-res Image Clustering Approach #6: 'METHLYATION CNMF' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #7: 'RPPA CNMF subtypes'

Table S55. Description of clustering approach #7: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	46	60	46	30	48

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.688 (logrank test), Q value = 0.9

Table S56. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	229	170	0.1 - 120.6 (10.9)
subtype1	46	35	0.2 - 47.6 (9.1)
subtype2	60	48	0.1 - 53.2 (12.5)
subtype3	45	32	0.5 - 115.9 (8.3)
subtype4	30	21	0.2 - 43.5 (10.6)
subtype5	48	34	0.1 - 120.6 (10.9)

Figure S49. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.856 (Kruskal-Wallis (anova)), Q value = 0.93

Table S57. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	230	59.4 (14.3)
subtype1	46	59.4 (12.9)
subtype2	60	59.5 (12.9)
subtype3	46	56.9 (17.3)
subtype4	30	60.1 (15.3)
subtype5	48	61.4 (13.5)

Figure S50. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.0616 (Fisher's exact test), Q value = 0.3

Table S58. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	91	139
subtype1	13	33
subtype2	19	41
subtype3	19	27
subtype4	14	16
subtype5	26	22

Figure S51. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RPPA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.767 (Fisher's exact test), Q value = 0.91

Table S59. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	29	186
subtype1	6	39
subtype2	5	48
subtype3	8	35
subtype4	4	23
subtype5	6	41

Figure S52. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'RPPA CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.988 (Kruskal-Wallis (anova)), Q value = 0.99

Table S60. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	176	75.5 (15.8)
subtype1	35	74.0 (20.3)
subtype2	46	75.4 (12.4)
subtype3	35	75.7 (17.7)
subtype4	22	75.9 (15.9)
subtype5	38	76.3 (13.4)

Figure S53. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.896 (Fisher's exact test), Q value = 0.93

Table S61. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	17	3	210
subtype1	3	0	43
subtype2	3	2	55
subtype3	4	0	42
subtype4	2	0	28
subtype5	5	1	42

Figure S54. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.987 (Fisher's exact test), Q value = 0.99

Table S62. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	4	27	183
subtype1	0	6	37
subtype2	1	8	47
subtype3	1	4	36
subtype4	1	3	23
subtype5	1	6	40

Figure S55. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.668 (Fisher's exact test), Q value = 0.9

Table S63. Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	194
subtype1	1	40
subtype2	2	48
subtype3	0	40
subtype4	0	25
subtype5	0	41

Figure S56. Get High-res Image Clustering Approach #7: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #8: 'RPPA cHierClus subtypes'

Table S64. Description of clustering approach #8: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	93	76	61

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.0238 (logrank test), Q value = 0.21

Table S65. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	229	170	0.1 - 120.6 (10.9)
subtype1	93	74	0.2 - 53.2 (8.4)
subtype2	75	50	0.1 - 115.9 (9.7)
subtype3	61	46	0.1 - 120.6 (13.9)

Figure S57. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0447 (Kruskal-Wallis (anova)), Q value = 0.26

Table S66. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	230	59.4 (14.3)
subtype1	93	61.2 (13.2)
subtype2	76	55.2 (16.7)
subtype3	61	62.1 (11.2)

Figure S58. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.236 (Fisher's exact test), Q value = 0.53

Table S67. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	91	139
subtype1	34	59
subtype2	36	40
subtype3	21	40

Figure S59. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RPPA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.647 (Fisher's exact test), Q value = 0.9

Table S68. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	29	186
subtype1	14	73
subtype2	9	62
subtype3	6	51

Figure S60. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'RPPA cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.068 (Kruskal-Wallis (anova)), Q value = 0.3

Table S69. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	176	75.5 (15.8)
subtype1	69	72.3 (17.1)
subtype2	61	78.5 (15.6)
subtype3	46	76.1 (13.4)

Figure S61. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.732 (Fisher's exact test), Q value = 0.91

Table S70. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	17	3	210
subtype1	8	2	83
subtype2	6	0	70
subtype3	3	1	57

Figure S62. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.42 (Fisher's exact test), Q value = 0.75

Table S71. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	4	27	183
subtype1	0	13	73
subtype2	2	7	60
subtype3	2	7	50

Figure S63. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.781 (Fisher's exact test), Q value = 0.91

Table S72. Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	194
subtype1	2	75
subtype2	1	65
subtype3	0	54

Figure S64. Get High-res Image Clustering Approach #8: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #9: 'RNAseq CNMF subtypes'

Table S73. Description of clustering approach #9: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	47	70	35

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.45 (logrank test), Q value = 0.78

Table S74. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	152	118	0.2 - 88.1 (10.9)
subtype1	47	34	0.2 - 88.1 (10.9)
subtype2	70	55	0.9 - 69.9 (10.9)
subtype3	35	29	0.2 - 46.9 (10.8)

Figure S65. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0964 (Kruskal-Wallis (anova)), Q value = 0.34

Table S75. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	152	59.7 (13.5)
subtype1	47	55.6 (16.6)
subtype2	70	62.0 (10.6)
subtype3	35	60.9 (13.3)

Figure S66. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.0701 (Fisher's exact test), Q value = 0.3

Table S76. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	53	99
subtype1	14	33
subtype2	31	39
subtype3	8	27

Figure S67. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.706 (Fisher's exact test), Q value = 0.91

Table S77. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	23	121
subtype1	6	38
subtype2	10	56
subtype3	7	27

Figure S68. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'RNAseq CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.41 (Kruskal-Wallis (anova)), Q value = 0.74

Table S78. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	117	75.7 (14.8)
subtype1	38	75.0 (15.2)
subtype2	53	74.5 (13.8)
subtype3	26	79.2 (16.0)

Figure S69. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.791 (Fisher's exact test), Q value = 0.91

Table S79. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	1	1	150
subtype1	1	0	46
subtype2	0	1	69
subtype3	0	0	35

Figure S70. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.792 (Fisher's exact test), Q value = 0.91

Table S80. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	10	136
subtype1	2	3	42
subtype2	1	5	63
subtype3	2	2	31

Figure S71. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.792 (Fisher's exact test), Q value = 0.91

Table S81. Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	125
subtype1	1	44
subtype2	2	51
subtype3	0	30

Figure S72. Get High-res Image Clustering Approach #9: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'ETHNICITY'

Clustering Approach #10: 'RNAseq cHierClus subtypes'

Table S82. Description of clustering approach #10: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	32	96	24

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0741 (logrank test), Q value = 0.3

Table S83. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	152	118	0.2 - 88.1 (10.9)
subtype1	32	22	0.2 - 88.1 (13.8)
subtype2	96	76	0.4 - 69.9 (10.1)
subtype3	24	20	0.2 - 46.9 (8.9)

Figure S73. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0784 (Kruskal-Wallis (anova)), Q value = 0.3

Table S84. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	152	59.7 (13.5)
subtype1	32	54.2 (18.3)
subtype2	96	61.6 (11.4)
subtype3	24	59.6 (12.5)

Figure S74. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.145 (Fisher's exact test), Q value = 0.41

Table S85. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	53	99
subtype1	9	23
subtype2	39	57
subtype3	5	19

Figure S75. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.898 (Fisher's exact test), Q value = 0.93

Table S86. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	23	121
subtype1	4	27
subtype2	15	74
subtype3	4	20

Figure S76. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'RADIATION_THERAPY'

'RNAseq cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0366 (Kruskal-Wallis (anova)), Q value = 0.24

Table S87. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	117	75.7 (14.8)
subtype1	26	78.1 (11.7)
subtype2	75	73.3 (14.6)
subtype3	16	83.1 (17.4)

Figure S77. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'KARNOFSKY_PERFORMANCE_SCORE'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.602 (Fisher's exact test), Q value = 0.88

Table S88. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

nPatients	GLIOBLASTOMA MULTIFORME (GBM)	TREATED PRIMARY GBM	UNTREATED PRIMARY (DE NOVO) GBM
ALL	1	1	150
subtype1	1	0	31
subtype2	0	1	95
subtype3	0	0	24

Figure S78. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.577 (Fisher's exact test), Q value = 0.88

Table S89. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	10	136
subtype1	2	1	29
subtype2	2	8	85
subtype3	1	1	22

Figure S79. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.738 (Fisher's exact test), Q value = 0.91

Table S90. Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	125
subtype1	0	30
subtype2	3	75
subtype3	0	20

Figure S80. Get High-res Image Clustering Approach #10: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'ETHNICITY'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/GBM-TP/20125497/GBM-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/GBM-TP/19775175/GBM-TP.merged_data.txt
Number of patients = 593
Number of clustering approaches = 10
Number of selected clinical features = 8
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)

[2] Monti et al., Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data, Machine Learning 52(1):91-118 (2003)

[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)

Made with Nozzle