Correlation between aggregated molecular cancer subtypes and selected clinical features

Breast Invasive Carcinoma (Primary solid tumor)

17 October 2014 | analyses__2014_10_17

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/C1JW8CQ0

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 12 different clustering approaches and 12 clinical features across 1043 patients, 73 significant findings detected with P value < 0.05 and Q value < 0.25.

CNMF clustering analysis on array-based mRNA expression data identified 8 subtypes that correlate to 'AGE', 'PATHOLOGY.N.STAGE', 'HISTOLOGICAL.TYPE', and 'RACE'.
Consensus hierarchical clustering analysis on array-based mRNA expression data identified 5 subtypes that correlate to 'AGE', 'HISTOLOGICAL.TYPE', and 'RACE'.
5 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'PATHOLOGY.T.STAGE', 'GENDER', 'HISTOLOGICAL.TYPE', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.
5 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'AGE', 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'HISTOLOGICAL.TYPE', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.
CNMF clustering analysis on RPPA data identified 3 subtypes that correlate to 'PATHOLOGY.T.STAGE' and 'HISTOLOGICAL.TYPE'.
Consensus hierarchical clustering analysis on RPPA data identified 5 subtypes that correlate to 'AGE', 'PATHOLOGY.T.STAGE', 'HISTOLOGICAL.TYPE', and 'RACE'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'AGE', 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'HISTOLOGICAL.TYPE', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 6 subtypes that correlate to 'Time to Death', 'AGE', 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'GENDER', 'HISTOLOGICAL.TYPE', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'PATHOLOGY.M.STAGE', 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', and 'RACE'.
6 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'Time to Death', 'AGE', 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'PATHOLOGY.M.STAGE', 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.
5 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'PATHOLOGY.M.STAGE', 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', and 'RACE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'NEOPLASM.DISEASESTAGE', 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'PATHOLOGY.M.STAGE', 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', 'NUMBER.OF.LYMPH.NODES', and 'RACE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 12 different clustering approaches and 12 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 73 significant findings detected.


Clinical Features	Time to Death	AGE	NEOPLASM DISEASESTAGE	PATHOLOGY T STAGE	PATHOLOGY N STAGE	PATHOLOGY M STAGE	GENDER	HISTOLOGICAL TYPE	RADIATIONS RADIATION REGIMENINDICATION	NUMBER OF LYMPH NODES	RACE	ETHNICITY
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	Fisher's exact test	Fisher's exact test	Kruskal-Wallis (anova)	Fisher's exact test	Fisher's exact test
mRNA CNMF subtypes	0.0484 (1.00)	3.15e-08 (4.51e-06)	0.0406 (1.00)	0.0363 (1.00)	0.00119 (0.0988)	0.61 (1.00)	0.0276 (1.00)	1e-05 (0.0014)	0.311 (1.00)	0.00993 (0.655)	0.00246 (0.184)	0.284 (1.00)
mRNA cHierClus subtypes	0.112 (1.00)	5.4e-10 (7.78e-08)	0.0456 (1.00)	0.00418 (0.297)	0.00531 (0.367)	0.52 (1.00)	0.101 (1.00)	1e-05 (0.0014)	0.0105 (0.681)	0.0179 (1.00)	0.00134 (0.109)	0.76 (1.00)
Copy Number Ratio CNMF subtypes	0.132 (1.00)	0.00524 (0.367)	0.00674 (0.452)	1e-05 (0.0014)	0.00532 (0.367)	0.0179 (1.00)	3e-05 (0.00333)	1e-05 (0.0014)	0.202 (1.00)	0.00264 (0.193)	0.00011 (0.011)	0.872 (1.00)
METHLYATION CNMF	0.015 (0.918)	5.53e-06 (0.00078)	0.00029 (0.027)	0.0006 (0.0534)	0.00018 (0.0176)	0.701 (1.00)	0.272 (1.00)	1e-05 (0.0014)	0.809 (1.00)	0.000676 (0.0588)	1e-05 (0.0014)	0.787 (1.00)
RPPA CNMF subtypes	0.117 (1.00)	0.0119 (0.749)	0.0335 (1.00)	0.00218 (0.17)	0.154 (1.00)	0.386 (1.00)	0.0567 (1.00)	3e-05 (0.00333)	0.728 (1.00)	0.551 (1.00)	0.0576 (1.00)	1 (1.00)
RPPA cHierClus subtypes	0.33 (1.00)	0.00228 (0.173)	0.0187 (1.00)	7e-05 (0.00714)	0.0896 (1.00)	0.364 (1.00)	0.0842 (1.00)	0.00021 (0.0202)	0.0696 (1.00)	0.0409 (1.00)	0.00122 (0.1)	0.967 (1.00)
RNAseq CNMF subtypes	0.0261 (1.00)	2.87e-05 (0.00322)	1e-05 (0.0014)	1e-05 (0.0014)	0.00062 (0.0546)	0.0252 (1.00)	0.0142 (0.878)	1e-05 (0.0014)	0.622 (1.00)	0.00224 (0.172)	1e-05 (0.0014)	0.366 (1.00)
RNAseq cHierClus subtypes	3.45e-05 (0.00365)	9.4e-07 (0.000134)	3e-05 (0.00333)	1e-05 (0.0014)	1e-05 (0.0014)	0.0197 (1.00)	0.00146 (0.117)	1e-05 (0.0014)	0.219 (1.00)	6.59e-05 (0.00678)	1e-05 (0.0014)	0.558 (1.00)
MIRSEQ CNMF	0.0223 (1.00)	0.244 (1.00)	1e-05 (0.0014)	4e-05 (0.0042)	0.00086 (0.0731)	5e-05 (0.0052)	0.182 (1.00)	1e-05 (0.0014)	0.00028 (0.0266)	0.0209 (1.00)	1e-05 (0.0014)	0.411 (1.00)
MIRSEQ CHIERARCHICAL	0.000316 (0.0291)	0.00268 (0.193)	0.0005 (0.045)	0.00028 (0.0266)	7e-05 (0.00714)	0.00108 (0.0907)	0.0678 (1.00)	1e-05 (0.0014)	0.00069 (0.0593)	0.00013 (0.0128)	1e-05 (0.0014)	0.365 (1.00)
MIRseq Mature CNMF subtypes	0.229 (1.00)	0.0226 (1.00)	0.00041 (0.0373)	0.00018 (0.0176)	0.00151 (0.119)	1e-05 (0.0014)	0.413 (1.00)	1e-05 (0.0014)	1e-05 (0.0014)	0.0161 (0.965)	1e-05 (0.0014)	0.431 (1.00)
MIRseq Mature cHierClus subtypes	0.0309 (1.00)	0.0105 (0.681)	1e-05 (0.0014)	2e-05 (0.00228)	3e-05 (0.00333)	2e-05 (0.00228)	0.805 (1.00)	1e-05 (0.0014)	1e-05 (0.0014)	0.0026 (0.192)	3e-05 (0.00333)	0.602 (1.00)

Clustering Approach #1: 'mRNA CNMF subtypes'

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

Cluster Labels	1	2	3	4	5	6	7	8
Number of samples	21	39	121	101	110	74	20	40

'mRNA CNMF subtypes' versus 'Time to Death'

P value = 0.0484 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	511	73	0.1 - 234.3 (29.0)
subtype1	20	3	0.3 - 92.0 (15.3)
subtype2	39	3	1.4 - 157.4 (43.6)
subtype3	116	19	0.2 - 188.7 (29.1)
subtype4	100	13	0.2 - 211.6 (30.2)
subtype5	108	11	0.3 - 234.3 (27.7)
subtype6	70	15	0.1 - 189.0 (31.0)
subtype7	19	3	1.0 - 97.5 (44.8)
subtype8	39	6	0.3 - 112.4 (27.5)

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

'mRNA CNMF subtypes' versus 'AGE'

P value = 3.15e-08 (Kruskal-Wallis (anova)), Q value = 4.5e-06

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

	nPatients	Mean (Std.Dev)
ALL	520	58.1 (13.2)
subtype1	20	60.7 (13.6)
subtype2	38	48.5 (10.6)
subtype3	119	58.3 (14.3)
subtype4	101	53.9 (12.5)
subtype5	110	62.6 (12.5)
subtype6	72	59.1 (12.3)
subtype7	20	60.4 (9.9)
subtype8	40	60.6 (12.0)

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

'mRNA CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0406 (Fisher's exact test), Q value = 1

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

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	46	37	6	185	110	77	15	19	14	1	16
subtype1	1	0	0	8	3	5	1	0	2	0	1
subtype2	1	3	0	12	10	8	0	2	1	1	1
subtype3	8	6	4	36	34	20	6	3	1	0	3
subtype4	8	10	0	45	19	9	1	4	3	0	2
subtype5	14	14	0	43	15	10	5	2	3	0	4
subtype6	4	2	0	21	19	15	2	6	3	0	2
subtype7	4	1	0	6	4	4	0	0	0	0	1
subtype8	6	1	2	14	6	6	0	2	1	0	2

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

'mRNA CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.0363 (Fisher's exact test), Q value = 1

Table S5. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	133	311	59	20
subtype1	6	13	0	2
subtype2	8	21	9	0
subtype3	25	75	14	6
subtype4	23	67	8	2
subtype5	40	55	9	6
subtype6	11	49	12	2
subtype7	7	10	3	0
subtype8	13	21	4	2

Figure S4. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'mRNA CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.00119 (Fisher's exact test), Q value = 0.099

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

nPatients	N0	N1	N2	N3
ALL	255	170	61	29
subtype1	6	6	7	1
subtype2	14	16	5	3
subtype3	49	50	14	4
subtype4	64	24	8	5
subtype5	66	31	7	5
subtype6	24	26	14	8
subtype7	11	5	3	0
subtype8	21	12	3	3

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

'mRNA CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.61 (Fisher's exact test), Q value = 1

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

nPatients	CM0 (I+)	M0	M1	MX
ALL	2	495	15	14
subtype1	0	19	2	0
subtype2	1	37	1	0
subtype3	0	115	2	4
subtype4	0	97	3	1
subtype5	1	101	3	5
subtype6	0	70	3	1
subtype7	0	19	0	1
subtype8	0	37	1	2

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

'mRNA CNMF subtypes' versus 'GENDER'

P value = 0.0276 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	520	6
subtype1	20	1
subtype2	39	0
subtype3	117	4
subtype4	101	0
subtype5	110	0
subtype6	74	0
subtype7	19	1
subtype8	40	0

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

'mRNA CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

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

nPatients	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	448	41	1	12	2	21
subtype1	20	0	0	0	0	1
subtype2	28	8	0	3	0	0
subtype3	113	1	0	1	1	5
subtype4	92	1	1	0	0	6
subtype5	85	14	0	5	1	5
subtype6	67	5	0	0	0	2
subtype7	19	1	0	0	0	0
subtype8	24	11	0	3	0	2

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

'mRNA CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.311 (Fisher's exact test), Q value = 1

Table S10. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	168	358
subtype1	3	18
subtype2	14	25
subtype3	35	86
subtype4	41	60
subtype5	33	77
subtype6	22	52
subtype7	8	12
subtype8	12	28

Figure S9. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'mRNA CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.00993 (Kruskal-Wallis (anova)), Q value = 0.66

Table S11. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	408	1.8 (3.5)
subtype1	14	1.4 (2.2)
subtype2	34	1.9 (2.4)
subtype3	86	1.6 (2.3)
subtype4	93	1.4 (3.2)
subtype5	84	1.8 (3.9)
subtype6	47	2.7 (3.5)
subtype7	15	1.2 (2.5)
subtype8	35	2.7 (6.3)

Figure S10. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'mRNA CNMF subtypes' versus 'RACE'

P value = 0.00246 (Fisher's exact test), Q value = 0.18

Table S12. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	34	40	361
subtype1	0	2	0	14
subtype2	0	4	3	28
subtype3	0	10	12	74
subtype4	0	6	12	75
subtype5	0	2	5	81
subtype6	1	9	2	43
subtype7	0	1	5	11
subtype8	0	0	1	35

Figure S11. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #11: 'RACE'

'mRNA CNMF subtypes' versus 'ETHNICITY'

P value = 0.284 (Fisher's exact test), Q value = 1

Table S13. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	7	372
subtype1	0	14
subtype2	1	32
subtype3	0	83
subtype4	1	83
subtype5	2	70
subtype6	2	46
subtype7	1	14
subtype8	0	30

Figure S12. Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #2: 'mRNA cHierClus subtypes'

Table S14. Description of clustering approach #2: 'mRNA cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	102	156	118	92	58

'mRNA cHierClus subtypes' versus 'Time to Death'

P value = 0.112 (logrank test), Q value = 1

Table S15. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	511	73	0.1 - 234.3 (29.0)
subtype1	98	16	0.2 - 234.3 (30.8)
subtype2	155	17	0.3 - 220.9 (31.8)
subtype3	112	21	0.1 - 189.0 (25.0)
subtype4	91	12	0.3 - 211.6 (31.7)
subtype5	55	7	0.3 - 129.6 (23.9)

Figure S13. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'mRNA cHierClus subtypes' versus 'AGE'

P value = 5.4e-10 (Kruskal-Wallis (anova)), Q value = 7.8e-08

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

	nPatients	Mean (Std.Dev)
ALL	520	58.1 (13.2)
subtype1	100	59.8 (13.1)
subtype2	154	55.2 (13.0)
subtype3	116	59.6 (12.2)
subtype4	92	53.5 (12.6)
subtype5	58	66.9 (11.9)

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

'mRNA cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0456 (Fisher's exact test), Q value = 1

Table S17. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	46	37	6	185	110	77	15	19	14	1	16
subtype1	6	5	2	33	29	16	5	2	1	0	3
subtype2	20	12	4	56	24	23	1	6	3	1	6
subtype3	6	4	0	36	27	24	5	8	6	0	2
subtype4	8	9	0	40	18	8	1	3	3	0	2
subtype5	6	7	0	20	12	6	3	0	1	0	3

Figure S15. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.00418 (Fisher's exact test), Q value = 0.3

Table S18. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	133	311	59	20
subtype1	19	63	14	5
subtype2	54	79	19	3
subtype3	18	78	15	7
subtype4	22	60	8	1
subtype5	20	31	3	4

Figure S16. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.00531 (Fisher's exact test), Q value = 0.37

Table S19. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	255	170	61	29
subtype1	44	43	9	4
subtype2	81	47	15	9
subtype3	41	39	25	10
subtype4	58	23	7	4
subtype5	31	18	5	2

Figure S17. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.52 (Fisher's exact test), Q value = 1

Table S20. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	2	495	15	14
subtype1	0	96	2	4
subtype2	2	146	3	5
subtype3	0	111	6	1
subtype4	0	88	3	1
subtype5	0	54	1	3

Figure S18. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'mRNA cHierClus subtypes' versus 'GENDER'

P value = 0.101 (Fisher's exact test), Q value = 1

Table S21. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	520	6
subtype1	98	4
subtype2	155	1
subtype3	117	1
subtype4	92	0
subtype5	58	0

Figure S19. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'mRNA cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S22. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	448	41	1	12	2	21
subtype1	97	1	0	0	1	3
subtype2	113	29	0	7	0	7
subtype3	110	6	0	0	0	2
subtype4	84	1	1	0	0	5
subtype5	44	4	0	5	1	4

Figure S20. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'mRNA cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.0105 (Fisher's exact test), Q value = 0.68

Table S23. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	168	358
subtype1	28	74
subtype2	46	110
subtype3	29	89
subtype4	40	52
subtype5	25	33

Figure S21. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'mRNA cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0179 (Kruskal-Wallis (anova)), Q value = 1

Table S24. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	408	1.8 (3.5)
subtype1	70	1.4 (1.9)
subtype2	135	2.0 (4.4)
subtype3	74	2.4 (3.4)
subtype4	86	1.5 (3.3)
subtype5	43	1.5 (2.7)

Figure S22. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'mRNA cHierClus subtypes' versus 'RACE'

P value = 0.00134 (Fisher's exact test), Q value = 0.11

Table S25. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	34	40	361
subtype1	0	7	10	60
subtype2	0	6	8	128
subtype3	1	16	4	65
subtype4	0	3	12	70
subtype5	0	2	6	38

Figure S23. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #11: 'RACE'

'mRNA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.76 (Fisher's exact test), Q value = 1

Table S26. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	7	372
subtype1	0	66
subtype2	3	119
subtype3	2	74
subtype4	1	75
subtype5	1	38

Figure S24. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #3: 'Copy Number Ratio CNMF subtypes'

Table S27. Description of clustering approach #3: 'Copy Number Ratio CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	249	387	85	256	48

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

P value = 0.132 (logrank test), Q value = 1

Table S28. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	1005	119	0.0 - 234.3 (22.4)
subtype1	240	32	0.1 - 162.1 (22.0)
subtype2	380	40	0.0 - 234.3 (24.9)
subtype3	84	9	0.0 - 189.0 (20.6)
subtype4	253	30	0.0 - 211.6 (21.1)
subtype5	48	8	0.2 - 220.9 (21.2)

Figure S25. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.00524 (Kruskal-Wallis (anova)), Q value = 0.37

Table S29. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	1012	58.6 (13.2)
subtype1	242	59.3 (14.0)
subtype2	384	58.5 (13.1)
subtype3	85	62.0 (12.1)
subtype4	255	56.7 (12.9)
subtype5	46	61.5 (12.1)

Figure S26. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

'Copy Number Ratio CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.00674 (Fisher's exact test), Q value = 0.45

Table S30. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	83	80	9	5	342	234	2	142	29	63	18	1	16
subtype1	13	15	2	1	80	51	1	47	14	15	8	0	2
subtype2	43	44	6	3	106	89	1	49	6	27	5	0	8
subtype3	5	2	0	1	28	26	0	9	4	6	2	0	2
subtype4	18	16	1	0	108	57	0	31	5	12	3	1	3
subtype5	4	3	0	0	20	11	0	6	0	3	0	0	1

Figure S27. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S31. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	265	594	124	39
subtype1	49	159	26	15
subtype2	143	174	60	9
subtype3	9	57	12	7
subtype4	49	176	22	7
subtype5	15	28	4	1

Figure S28. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.00532 (Fisher's exact test), Q value = 0.37

Table S32. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	481	339	113	74
subtype1	96	85	41	19
subtype2	184	135	30	30
subtype3	36	32	6	9
subtype4	144	69	30	13
subtype5	21	18	6	3

Figure S29. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.0179 (Fisher's exact test), Q value = 1

Table S33. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	854	19	147
subtype1	1	216	9	23
subtype2	3	307	5	72
subtype3	0	70	2	13
subtype4	0	222	3	31
subtype5	1	39	0	8

Figure S30. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 3e-05 (Fisher's exact test), Q value = 0.0033

Table S34. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	1014	11
subtype1	239	10
subtype2	387	0
subtype3	84	1
subtype4	256	0
subtype5	48	0

Figure S31. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S35. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	754	175	6	29	15	44
subtype1	0	216	19	0	4	3	7
subtype2	0	226	113	0	15	11	22
subtype3	0	52	27	0	5	0	1
subtype4	1	226	7	5	4	0	12
subtype5	0	34	9	1	1	1	2

Figure S32. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.202 (Fisher's exact test), Q value = 1

Table S36. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	282	743
subtype1	74	175
subtype2	93	294
subtype3	20	65
subtype4	79	177
subtype5	16	32

Figure S33. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'Copy Number Ratio CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.00264 (Kruskal-Wallis (anova)), Q value = 0.19

Table S37. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	864	2.4 (4.7)
subtype1	187	2.7 (4.5)
subtype2	339	2.3 (4.6)
subtype3	71	3.0 (5.2)
subtype4	223	2.0 (4.8)
subtype5	44	2.9 (5.6)

Figure S34. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.00011 (Fisher's exact test), Q value = 0.011

Table S38. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	56	158	717
subtype1	0	20	29	159
subtype2	1	15	51	295
subtype3	0	2	6	67
subtype4	0	16	65	159
subtype5	0	3	7	37

Figure S35. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'RACE'

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

P value = 0.872 (Fisher's exact test), Q value = 1

Table S39. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	36	826
subtype1	6	193
subtype2	16	316
subtype3	2	64
subtype4	10	212
subtype5	2	41

Figure S36. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #4: 'METHLYATION CNMF'

Table S40. Description of clustering approach #4: 'METHLYATION CNMF'

Cluster Labels	1	2	3	4	5
Number of samples	129	195	172	86	119

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.015 (logrank test), Q value = 0.92

Table S41. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	696	80	0.0 - 234.3 (22.3)
subtype1	129	18	0.2 - 211.6 (22.3)
subtype2	193	19	0.2 - 234.3 (23.9)
subtype3	171	29	0.0 - 173.0 (21.2)
subtype4	86	4	0.4 - 157.4 (20.7)
subtype5	117	10	0.0 - 194.3 (28.4)

Figure S37. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'AGE'

P value = 5.53e-06 (Kruskal-Wallis (anova)), Q value = 0.00078

Table S42. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	690	58.0 (13.1)
subtype1	128	55.2 (12.4)
subtype2	192	58.3 (13.4)
subtype3	170	62.1 (12.7)
subtype4	83	55.7 (13.4)
subtype5	117	56.4 (12.3)

Figure S38. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

'METHLYATION CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.00029 (Fisher's exact test), Q value = 0.027

Table S43. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE X
ALL	53	59	4	4	224	163	2	112	18	48	8	5
subtype1	6	9	0	0	61	31	0	12	2	3	3	1
subtype2	14	15	3	2	53	56	0	31	6	12	1	2
subtype3	9	11	0	1	49	38	1	38	8	12	3	2
subtype4	6	12	0	1	19	19	0	14	1	13	1	0
subtype5	18	12	1	0	42	19	1	17	1	8	0	0

Figure S39. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'METHLYATION CNMF' versus 'PATHOLOGY.T.STAGE'

P value = 6e-04 (Fisher's exact test), Q value = 0.053

Table S44. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	186	401	91	21
subtype1	23	89	12	4
subtype2	55	110	24	5
subtype3	33	100	29	10
subtype4	25	47	13	1
subtype5	50	55	13	1

Figure S40. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY.N.STAGE'

P value = 0.00018 (Fisher's exact test), Q value = 0.018

Table S45. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	309	242	89	51
subtype1	77	35	12	4
subtype2	67	88	22	13
subtype3	69	55	32	12
subtype4	35	28	10	13
subtype5	61	36	13	9

Figure S41. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY.M.STAGE'

P value = 0.701 (Fisher's exact test), Q value = 1

Table S46. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	1	564	9	127
subtype1	0	106	3	20
subtype2	1	160	1	33
subtype3	0	135	4	33
subtype4	0	68	1	17
subtype5	0	95	0	24

Figure S42. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.272 (Fisher's exact test), Q value = 1

Table S47. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	693	8
subtype1	129	0
subtype2	192	3
subtype3	169	3
subtype4	84	2
subtype5	119	0

Figure S43. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

'METHLYATION CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S48. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	479	147	5	24	12	32
subtype1	1	106	4	3	2	0	12
subtype2	0	118	45	0	14	7	11
subtype3	0	126	35	0	3	3	5
subtype4	0	51	30	0	3	1	1
subtype5	0	78	33	2	2	1	3

Figure S44. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'METHLYATION CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.809 (Fisher's exact test), Q value = 1

Table S49. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	215	486
subtype1	44	85
subtype2	60	135
subtype3	54	118
subtype4	23	63
subtype5	34	85

Figure S45. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'METHLYATION CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.000676 (Kruskal-Wallis (anova)), Q value = 0.059

Table S50. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	649	2.6 (4.9)
subtype1	122	1.5 (2.9)
subtype2	182	2.5 (4.4)
subtype3	153	3.3 (5.5)
subtype4	79	4.0 (6.2)
subtype5	113	2.3 (5.3)

Figure S46. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S51. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	34	119	533
subtype1	0	5	45	76
subtype2	0	8	20	164
subtype3	0	12	26	130
subtype4	0	4	8	73
subtype5	1	5	20	90

Figure S47. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #11: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.787 (Fisher's exact test), Q value = 1

Table S52. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	35	601
subtype1	6	109
subtype2	10	167
subtype3	11	147
subtype4	2	74
subtype5	6	104

Figure S48. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #5: 'RPPA CNMF subtypes'

Table S53. Description of clustering approach #5: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	154	135	120

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.117 (logrank test), Q value = 1

Table S54. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	396	52	0.1 - 189.0 (29.9)
subtype1	147	25	0.1 - 186.5 (29.0)
subtype2	133	13	0.2 - 146.5 (26.2)
subtype3	116	14	0.3 - 189.0 (33.8)

Figure S49. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'AGE'

P value = 0.0119 (Kruskal-Wallis (anova)), Q value = 0.75

Table S55. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	407	57.9 (13.1)
subtype1	154	56.0 (13.3)
subtype2	134	60.3 (13.8)
subtype3	119	57.6 (11.6)

Figure S50. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RPPA CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0335 (Fisher's exact test), Q value = 1

Table S56. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE X
ALL	32	30	4	137	94	62	13	15	14	8
subtype1	7	8	1	55	36	24	2	10	8	3
subtype2	12	12	0	46	38	14	6	1	3	3
subtype3	13	10	3	36	20	24	5	4	3	2

Figure S51. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.00218 (Fisher's exact test), Q value = 0.17

Table S57. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	94	246	50	18
subtype1	22	106	22	4
subtype2	34	83	11	6
subtype3	38	57	17	8

Figure S52. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.154 (Fisher's exact test), Q value = 1

Table S58. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	193	132	52	25
subtype1	72	45	22	12
subtype2	64	53	10	6
subtype3	57	34	20	7

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

'RPPA CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.386 (Fisher's exact test), Q value = 1

Table S59. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	1	386	15	7
subtype1	0	145	8	1
subtype2	0	128	3	4
subtype3	1	113	4	2

Figure S54. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.0567 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	404	5
subtype1	154	0
subtype2	131	4
subtype3	119	1

Figure S55. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 3e-05 (Fisher's exact test), Q value = 0.0033

Table S61. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	354	30	1	8	2	14
subtype1	141	4	1	1	0	7
subtype2	119	5	0	4	2	5
subtype3	94	21	0	3	0	2

Figure S56. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'RPPA CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.728 (Fisher's exact test), Q value = 1

Table S62. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	144	265
subtype1	57	97
subtype2	44	91
subtype3	43	77

Figure S57. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.551 (Kruskal-Wallis (anova)), Q value = 1

Table S63. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	333	1.8 (3.5)
subtype1	118	2.0 (3.6)
subtype2	112	1.2 (2.0)
subtype3	103	2.4 (4.5)

Figure S58. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.0576 (Fisher's exact test), Q value = 1

Table S64. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	27	29	294
subtype1	1	16	14	98
subtype2	0	7	9	101
subtype3	0	4	6	95

Figure S59. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

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

Table S65. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	297
subtype1	2	114
subtype2	2	98
subtype3	2	85

Figure S60. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #6: 'RPPA cHierClus subtypes'

Table S66. Description of clustering approach #6: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	64	111	93	109	32

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.33 (logrank test), Q value = 1

Table S67. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	396	52	0.1 - 189.0 (29.9)
subtype1	62	8	0.3 - 129.7 (28.4)
subtype2	109	11	0.2 - 173.0 (31.0)
subtype3	91	13	0.3 - 189.0 (31.7)
subtype4	104	15	0.2 - 129.6 (27.8)
subtype5	30	5	0.1 - 76.8 (27.9)

Figure S61. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'AGE'

P value = 0.00228 (Kruskal-Wallis (anova)), Q value = 0.17

Table S68. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	407	57.9 (13.1)
subtype1	63	61.9 (13.8)
subtype2	110	57.7 (11.8)
subtype3	93	54.2 (12.9)
subtype4	109	59.3 (13.2)
subtype5	32	56.5 (13.9)

Figure S62. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RPPA cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0187 (Fisher's exact test), Q value = 1

Table S69. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE X
ALL	32	30	4	137	94	62	13	15	14	8
subtype1	9	6	1	21	12	5	5	0	2	3
subtype2	11	13	3	31	22	20	3	3	3	2
subtype3	4	6	0	43	20	11	1	3	3	2
subtype4	7	5	0	32	30	22	4	4	4	1
subtype5	1	0	0	10	10	4	0	5	2	0

Figure S63. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 7e-05 (Fisher's exact test), Q value = 0.0071

Table S70. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	94	246	50	18
subtype1	22	32	3	6
subtype2	38	53	14	6
subtype3	14	68	10	1
subtype4	15	73	18	3
subtype5	5	20	5	2

Figure S64. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.0896 (Fisher's exact test), Q value = 1

Table S71. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	193	132	52	25
subtype1	35	19	5	3
subtype2	50	37	17	5
subtype3	53	25	11	4
subtype4	47	37	16	7
subtype5	8	14	3	6

Figure S65. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.364 (Fisher's exact test), Q value = 1

Table S72. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	1	386	15	7
subtype1	0	58	2	4
subtype2	1	104	4	2
subtype3	0	89	3	1
subtype4	0	105	4	0
subtype5	0	30	2	0

Figure S66. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.0842 (Fisher's exact test), Q value = 1

Table S73. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	404	5
subtype1	63	1
subtype2	111	0
subtype3	93	0
subtype4	105	4
subtype5	32	0

Figure S67. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.00021 (Fisher's exact test), Q value = 0.02

Table S74. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	354	30	1	8	2	14
subtype1	59	1	0	3	0	1
subtype2	90	17	0	3	0	1
subtype3	86	0	1	0	0	6
subtype4	90	10	0	2	2	5
subtype5	29	2	0	0	0	1

Figure S68. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'RPPA cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.0696 (Fisher's exact test), Q value = 1

Table S75. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	144	265
subtype1	24	40
subtype2	42	69
subtype3	41	52
subtype4	29	80
subtype5	8	24

Figure S69. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0409 (Kruskal-Wallis (anova)), Q value = 1

Table S76. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	333	1.8 (3.5)
subtype1	49	0.9 (1.9)
subtype2	98	2.3 (4.3)
subtype3	82	1.8 (3.6)
subtype4	82	1.6 (2.7)
subtype5	22	3.1 (4.2)

Figure S70. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.00122 (Fisher's exact test), Q value = 0.1

Table S77. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	27	29	294
subtype1	0	3	4	46
subtype2	0	2	4	94
subtype3	0	7	13	65
subtype4	0	9	7	72
subtype5	1	6	1	17

Figure S71. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.967 (Fisher's exact test), Q value = 1

Table S78. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	297
subtype1	1	47
subtype2	2	78
subtype3	1	77
subtype4	2	71
subtype5	0	24

Figure S72. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #7: 'RNAseq CNMF subtypes'

Table S79. Description of clustering approach #7: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	524	292	223

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.0261 (logrank test), Q value = 1

Table S80. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	1019	122	0.0 - 234.3 (22.8)
subtype1	509	66	0.0 - 234.3 (21.9)
subtype2	288	41	0.0 - 211.6 (22.6)
subtype3	222	15	0.0 - 194.3 (25.5)

Figure S73. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 2.87e-05 (Kruskal-Wallis (anova)), Q value = 0.0032

Table S81. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	1026	58.6 (13.2)
subtype1	515	60.4 (13.7)
subtype2	289	56.9 (12.5)
subtype3	222	56.8 (12.3)

Figure S74. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S82. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	83	83	10	5	347	238	2	142	29	63	18	1	17
subtype1	42	36	7	1	171	118	1	80	22	25	11	0	10
subtype2	16	25	0	1	127	65	0	30	6	12	5	0	4
subtype3	25	22	3	3	49	55	1	32	1	26	2	1	3

Figure S75. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S83. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	271	601	125	39
subtype1	131	312	53	27
subtype2	66	188	27	10
subtype3	74	101	45	2

Figure S76. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.00062 (Fisher's exact test), Q value = 0.055

Table S84. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	489	344	113	74
subtype1	229	182	68	32
subtype2	161	89	25	13
subtype3	99	73	20	29

Figure S77. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.0252 (Fisher's exact test), Q value = 1

Table S85. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	868	19	147
subtype1	1	444	12	67
subtype2	1	251	5	35
subtype3	3	173	2	45

Figure S78. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.0142 (Fisher's exact test), Q value = 0.88

Table S86. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	1028	11
subtype1	514	10
subtype2	292	0
subtype3	222	1

Figure S79. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S87. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	768	175	6	29	15	44
subtype1	0	405	64	0	17	14	24
subtype2	1	257	7	6	2	0	18
subtype3	0	106	104	0	10	1	2

Figure S80. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'RNAseq CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.622 (Fisher's exact test), Q value = 1

Table S88. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	285	754
subtype1	150	374
subtype2	79	213
subtype3	56	167

Figure S81. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.00224 (Kruskal-Wallis (anova)), Q value = 0.17

Table S89. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	876	2.4 (4.7)
subtype1	417	2.2 (4.0)
subtype2	250	1.9 (4.3)
subtype3	209	3.4 (6.1)

Figure S82. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S90. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	57	160	726
subtype1	0	31	62	364
subtype2	1	22	77	173
subtype3	0	4	21	189

Figure S83. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.366 (Fisher's exact test), Q value = 1

Table S91. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	36	832
subtype1	21	400
subtype2	7	245
subtype3	8	187

Figure S84. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #8: 'RNAseq cHierClus subtypes'

Table S92. Description of clustering approach #8: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3	4	5	6
Number of samples	214	158	290	139	181	57

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 3.45e-05 (logrank test), Q value = 0.0037

Table S93. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	1019	122	0.0 - 234.3 (22.8)
subtype1	209	26	0.1 - 188.7 (22.8)
subtype2	154	30	0.0 - 189.0 (20.9)
subtype3	286	20	0.0 - 234.3 (26.8)
subtype4	135	14	0.2 - 220.9 (20.1)
subtype5	180	22	0.2 - 211.6 (23.3)
subtype6	55	10	0.0 - 100.7 (20.1)

Figure S85. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 9.4e-07 (Kruskal-Wallis (anova)), Q value = 0.00013

Table S94. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	1026	58.6 (13.2)
subtype1	208	59.2 (13.8)
subtype2	155	61.8 (12.7)
subtype3	289	56.9 (12.8)
subtype4	138	62.0 (13.5)
subtype5	181	55.2 (12.2)
subtype6	55	59.7 (13.0)

Figure S86. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 3e-05 (Fisher's exact test), Q value = 0.0033

Table S95. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	83	83	10	5	347	238	2	142	29	63	18	1	17
subtype1	16	13	6	1	66	56	1	33	8	7	3	0	4
subtype2	7	7	0	1	49	34	0	31	8	11	7	0	3
subtype3	33	23	3	3	75	71	0	41	3	31	2	1	4
subtype4	14	18	1	0	50	27	0	16	5	3	2	0	3
subtype5	12	16	0	0	84	39	0	15	4	5	3	0	2
subtype6	1	6	0	0	23	11	1	6	1	6	1	0	1

Figure S87. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S96. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	271	601	125	39
subtype1	55	126	22	10
subtype2	23	107	18	10
subtype3	87	142	56	4
subtype4	48	75	9	7
subtype5	39	119	16	6
subtype6	19	32	4	2

Figure S88. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S97. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	489	344	113	74
subtype1	86	87	24	10
subtype2	62	45	31	13
subtype3	130	99	25	33
subtype4	75	45	13	5
subtype5	114	48	13	6
subtype6	22	20	7	7

Figure S89. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.0197 (Fisher's exact test), Q value = 1

Table S98. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	868	19	147
subtype1	0	189	4	21
subtype2	0	130	7	21
subtype3	4	226	2	58
subtype4	0	116	2	21
subtype5	1	157	3	20
subtype6	0	50	1	6

Figure S90. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.00146 (Fisher's exact test), Q value = 0.12

Table S99. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	1028	11
subtype1	206	8
subtype2	156	2
subtype3	290	0
subtype4	138	1
subtype5	181	0
subtype6	57	0

Figure S91. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S100. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	768	175	6	29	15	44
subtype1	0	183	13	0	6	2	10
subtype2	0	135	14	0	4	0	5
subtype3	0	152	121	1	9	1	6
subtype4	0	87	21	0	9	12	10
subtype5	1	159	1	5	1	0	13
subtype6	0	52	5	0	0	0	0

Figure S92. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'RNAseq cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.219 (Fisher's exact test), Q value = 1

Table S101. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	285	754
subtype1	67	147
subtype2	33	125
subtype3	76	214
subtype4	38	101
subtype5	57	124
subtype6	14	43

Figure S93. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 6.59e-05 (Kruskal-Wallis (anova)), Q value = 0.0068

Table S102. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	876	2.4 (4.7)
subtype1	164	2.0 (3.4)
subtype2	113	3.2 (5.1)
subtype3	268	3.0 (5.6)
subtype4	122	1.7 (3.4)
subtype5	167	1.3 (2.9)
subtype6	42	4.1 (8.0)

Figure S94. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S103. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	57	160	726
subtype1	0	18	27	140
subtype2	0	14	15	103
subtype3	0	8	32	235
subtype4	0	2	14	112
subtype5	0	5	60	108
subtype6	1	10	12	28

Figure S95. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.558 (Fisher's exact test), Q value = 1

Table S104. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	36	832
subtype1	8	166
subtype2	3	121
subtype3	13	232
subtype4	6	112
subtype5	6	153
subtype6	0	48

Figure S96. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #9: 'MIRSEQ CNMF'

Table S105. Description of clustering approach #9: 'MIRSEQ CNMF'

Cluster Labels	1	2	3
Number of samples	320	226	477

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0223 (logrank test), Q value = 1

Table S106. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	1003	119	0.0 - 234.3 (22.3)
subtype1	314	47	0.0 - 211.6 (21.6)
subtype2	225	16	0.0 - 194.3 (20.7)
subtype3	464	56	0.0 - 234.3 (24.0)

Figure S97. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 0.244 (Kruskal-Wallis (anova)), Q value = 1

Table S107. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	1010	58.6 (13.2)
subtype1	316	57.8 (13.2)
subtype2	224	58.5 (13.1)
subtype3	470	59.2 (13.3)

Figure S98. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

'MIRSEQ CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S108. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	83	82	10	5	342	233	2	140	28	62	17	1	17
subtype1	18	18	1	0	140	68	0	43	7	13	6	0	5
subtype2	14	24	3	4	58	54	1	31	5	29	2	0	1
subtype3	51	40	6	1	144	111	1	66	16	20	9	1	11

Figure S99. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.T.STAGE'

P value = 4e-05 (Fisher's exact test), Q value = 0.0042

Table S109. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	270	589	122	39
subtype1	61	218	29	11
subtype2	64	115	42	5
subtype3	145	256	51	23

Figure S100. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.N.STAGE'

P value = 0.00086 (Fisher's exact test), Q value = 0.073

Table S110. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	483	338	111	72
subtype1	172	91	38	15
subtype2	99	75	19	30
subtype3	212	172	54	27

Figure S101. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.M.STAGE'

P value = 5e-05 (Fisher's exact test), Q value = 0.0052

Table S111. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	852	18	148
subtype1	0	274	6	40
subtype2	2	165	2	57
subtype3	3	413	10	51

Figure S102. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.182 (Fisher's exact test), Q value = 1

Table S112. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	1012	11
subtype1	316	4
subtype2	226	0
subtype3	470	7

Figure S103. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S113. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	754	174	6	28	15	44
subtype1	1	283	8	6	3	5	13
subtype2	0	101	112	0	4	3	6
subtype3	0	370	54	0	21	7	25

Figure S104. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'MIRSEQ CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.00028 (Fisher's exact test), Q value = 0.027

Table S114. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	278	745
subtype1	97	223
subtype2	38	188
subtype3	143	334

Figure S105. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0209 (Kruskal-Wallis (anova)), Q value = 1

Table S115. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	866	2.4 (4.7)
subtype1	268	2.0 (4.1)
subtype2	203	3.6 (6.4)
subtype3	395	2.1 (3.9)

Figure S106. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S116. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	57	159	718
subtype1	0	23	80	193
subtype2	0	8	40	171
subtype3	1	26	39	354

Figure S107. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #11: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.411 (Fisher's exact test), Q value = 1

Table S117. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	36	825
subtype1	8	266
subtype2	9	202
subtype3	19	357

Figure S108. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #10: 'MIRSEQ CHIERARCHICAL'

Table S118. Description of clustering approach #10: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3	4	5	6
Number of samples	231	152	299	126	85	130

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.000316 (logrank test), Q value = 0.029

Table S119. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	1003	119	0.0 - 234.3 (22.3)
subtype1	222	39	0.0 - 146.5 (23.3)
subtype2	150	22	0.0 - 194.3 (20.0)
subtype3	298	26	0.0 - 170.2 (21.2)
subtype4	124	10	0.3 - 189.0 (24.2)
subtype5	80	5	0.2 - 234.3 (22.6)
subtype6	129	17	0.2 - 211.6 (24.4)

Figure S109. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.00268 (Kruskal-Wallis (anova)), Q value = 0.19

Table S120. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	1010	58.6 (13.2)
subtype1	226	58.7 (13.9)
subtype2	151	58.2 (12.8)
subtype3	294	59.4 (13.2)
subtype4	125	59.0 (11.6)
subtype5	84	62.2 (14.9)
subtype6	130	54.8 (12.3)

Figure S110. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM.DISEASESTAGE'

P value = 5e-04 (Fisher's exact test), Q value = 0.045

Table S121. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE TIS	STAGE X
ALL	83	82	10	5	342	233	2	140	28	62	17	1	17
subtype1	14	16	2	0	73	52	0	42	7	14	5	1	5
subtype2	13	9	1	2	53	33	0	20	3	11	6	0	1
subtype3	25	29	5	3	78	78	1	44	6	28	1	0	1
subtype4	17	9	1	0	45	23	1	11	5	5	3	0	6
subtype5	7	6	1	0	30	21	0	13	5	0	1	0	1
subtype6	7	13	0	0	63	26	0	10	2	4	1	0	3

Figure S111. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.T.STAGE'

P value = 0.00028 (Fisher's exact test), Q value = 0.027

Table S122. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	270	589	122	39
subtype1	52	139	29	10
subtype2	40	95	13	4
subtype3	92	147	54	5
subtype4	35	74	7	10
subtype5	24	46	9	6
subtype6	27	88	10	4

Figure S112. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.N.STAGE'

P value = 7e-05 (Fisher's exact test), Q value = 0.0071

Table S123. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	483	338	111	72
subtype1	90	83	33	18
subtype2	73	43	21	12
subtype3	129	107	30	29
subtype4	68	37	10	9
subtype5	38	36	9	0
subtype6	85	32	8	4

Figure S113. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.M.STAGE'

P value = 0.00108 (Fisher's exact test), Q value = 0.091

Table S124. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	852	18	148
subtype1	1	203	6	21
subtype2	1	126	6	19
subtype3	1	231	1	66
subtype4	2	107	3	14
subtype5	0	73	1	11
subtype6	0	112	1	17

Figure S114. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.0678 (Fisher's exact test), Q value = 1

Table S125. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	1012	11
subtype1	229	2
subtype2	149	3
subtype3	298	1
subtype4	124	2
subtype5	82	3
subtype6	130	0

Figure S115. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S126. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	754	174	6	28	15	44
subtype1	0	198	12	0	6	5	10
subtype2	0	133	9	0	1	0	9
subtype3	0	148	130	1	9	3	8
subtype4	0	94	21	0	7	0	4
subtype5	0	67	1	0	4	7	6
subtype6	1	114	1	5	1	0	7

Figure S116. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.00069 (Fisher's exact test), Q value = 0.059

Table S127. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	278	745
subtype1	82	149
subtype2	47	105
subtype3	55	244
subtype4	33	93
subtype5	24	61
subtype6	37	93

Figure S117. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.00013 (Kruskal-Wallis (anova)), Q value = 0.013

Table S128. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	866	2.4 (4.7)
subtype1	190	2.6 (4.1)
subtype2	128	2.9 (5.8)
subtype3	271	3.0 (5.6)
subtype4	101	1.8 (4.2)
subtype5	61	1.4 (1.9)
subtype6	115	1.1 (2.5)

Figure S118. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S129. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	57	159	718
subtype1	1	18	22	170
subtype2	0	11	22	104
subtype3	0	13	45	235
subtype4	0	6	12	89
subtype5	0	4	9	50
subtype6	0	5	49	70

Figure S119. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.365 (Fisher's exact test), Q value = 1

Table S130. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	36	825
subtype1	5	183
subtype2	5	125
subtype3	13	260
subtype4	5	94
subtype5	5	50
subtype6	3	113

Figure S120. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #11: 'MIRseq Mature CNMF subtypes'

Table S131. Description of clustering approach #11: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	179	131	84	101	119

'MIRseq Mature CNMF subtypes' versus 'Time to Death'

P value = 0.229 (logrank test), Q value = 1

Table S132. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	605	64	0.0 - 194.3 (20.4)
subtype1	173	24	0.1 - 189.0 (20.2)
subtype2	129	16	0.0 - 130.2 (22.1)
subtype3	83	3	0.1 - 109.3 (15.3)
subtype4	101	13	0.3 - 120.6 (23.7)
subtype5	119	8	0.0 - 194.3 (19.7)

Figure S121. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.0226 (Kruskal-Wallis (anova)), Q value = 1

Table S133. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	601	58.3 (13.2)
subtype1	171	58.4 (14.1)
subtype2	131	57.4 (13.2)
subtype3	84	61.3 (12.7)
subtype4	100	55.4 (12.1)
subtype5	115	59.5 (12.8)

Figure S122. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.00041 (Fisher's exact test), Q value = 0.037

Table S134. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE X
ALL	61	46	5	5	207	145	2	80	13	40	5	4
subtype1	20	17	1	1	64	39	0	26	4	5	1	1
subtype2	10	7	0	0	64	25	0	17	3	3	0	1
subtype3	5	8	4	1	20	19	0	11	1	13	2	0
subtype4	15	8	0	0	27	27	0	12	2	7	1	2
subtype5	11	6	0	3	32	35	2	14	3	12	1	0

Figure S123. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.00018 (Fisher's exact test), Q value = 0.018

Table S135. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	175	348	73	18
subtype1	59	101	13	6
subtype2	27	91	8	5
subtype3	19	45	18	2
subtype4	42	45	12	2
subtype5	28	66	22	3

Figure S124. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.00151 (Fisher's exact test), Q value = 0.12

Table S136. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	293	211	60	42
subtype1	85	66	19	6
subtype2	79	33	16	3
subtype3	33	29	6	13
subtype4	43	42	8	7
subtype5	53	41	11	13

Figure S125. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S137. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	481	5	123
subtype1	1	157	1	20
subtype2	0	106	0	25
subtype3	1	44	2	37
subtype4	3	77	1	20
subtype5	0	97	1	21

Figure S126. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.413 (Fisher's exact test), Q value = 1

Table S138. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	606	8
subtype1	175	4
subtype2	130	1
subtype3	82	2
subtype4	100	1
subtype5	119	0

Figure S127. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S139. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	420	137	5	15	13	23
subtype1	0	146	11	0	10	6	6
subtype2	1	118	2	5	1	0	4
subtype3	0	35	43	0	1	4	1
subtype4	0	62	29	0	2	1	7
subtype5	0	59	52	0	1	2	5

Figure S128. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S140. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	147	467
subtype1	62	117
subtype2	41	90
subtype3	4	80
subtype4	23	78
subtype5	17	102

Figure S129. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0161 (Kruskal-Wallis (anova)), Q value = 0.96

Table S141. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	552	2.5 (5.2)
subtype1	156	1.9 (3.9)
subtype2	122	1.4 (2.7)
subtype3	71	4.0 (6.6)
subtype4	97	2.7 (5.0)
subtype5	106	3.6 (7.2)

Figure S130. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S142. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	47	122	439
subtype1	20	10	149
subtype2	9	48	73
subtype3	0	32	51
subtype4	3	12	84
subtype5	15	20	82

Figure S131. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.431 (Fisher's exact test), Q value = 1

Table S143. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	26	560
subtype1	8	164
subtype2	3	122
subtype3	2	76
subtype4	7	87
subtype5	6	111

Figure S132. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'ETHNICITY'

Clustering Approach #12: 'MIRseq Mature cHierClus subtypes'

Table S144. Description of clustering approach #12: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	295	209	110

'MIRseq Mature cHierClus subtypes' versus 'Time to Death'

P value = 0.0309 (logrank test), Q value = 1

Table S145. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	605	64	0.0 - 194.3 (20.4)
subtype1	286	40	0.0 - 144.6 (21.9)
subtype2	209	11	0.0 - 194.3 (16.4)
subtype3	110	13	0.8 - 189.0 (22.7)

Figure S133. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.0105 (Kruskal-Wallis (anova)), Q value = 0.68

Table S146. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	601	58.3 (13.2)
subtype1	287	57.7 (13.8)
subtype2	205	60.3 (12.7)
subtype3	109	55.8 (12.3)

Figure S134. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S147. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE X
ALL	61	46	5	5	207	145	2	80	13	40	5	4
subtype1	27	21	2	1	121	64	0	42	7	7	1	1
subtype2	15	15	3	4	52	55	2	27	5	29	2	0
subtype3	19	10	0	0	34	26	0	11	1	4	2	3

Figure S135. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM.DISEASESTAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 2e-05 (Fisher's exact test), Q value = 0.0023

Table S148. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	175	348	73	18
subtype1	75	185	25	10
subtype2	53	109	42	5
subtype3	47	54	6	3

Figure S136. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 3e-05 (Fisher's exact test), Q value = 0.0033

Table S149. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	293	211	60	42
subtype1	155	94	35	7
subtype2	87	72	17	30
subtype3	51	45	8	5

Figure S137. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 2e-05 (Fisher's exact test), Q value = 0.0023

Table S150. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

nPatients	CM0 (I+)	M0	M1	MX
ALL	5	481	5	123
subtype1	1	251	1	42
subtype2	2	140	2	65
subtype3	2	90	2	16

Figure S138. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.M.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.805 (Fisher's exact test), Q value = 1

Table S151. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	606	8
subtype1	291	4
subtype2	207	2
subtype3	108	2

Figure S139. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S152. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

nPatients	INFILTRATING CARCINOMA NOS	INFILTRATING DUCTAL CARCINOMA	INFILTRATING LOBULAR CARCINOMA	MEDULLARY CARCINOMA	MIXED HISTOLOGY (PLEASE SPECIFY)	MUCINOUS CARCINOMA	OTHER SPECIFY
ALL	1	420	137	5	15	13	23
subtype1	1	245	13	5	9	9	13
subtype2	0	93	106	0	2	3	5
subtype3	0	82	18	0	4	1	5

Figure S140. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL.TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 1e-05 (Fisher's exact test), Q value = 0.0014

Table S153. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	147	467
subtype1	93	202
subtype2	19	190
subtype3	35	75

Figure S141. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0026 (Kruskal-Wallis (anova)), Q value = 0.19

Table S154. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	552	2.5 (5.2)
subtype1	264	1.8 (3.7)
subtype2	184	3.9 (6.9)
subtype3	104	2.1 (4.2)

Figure S142. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 3e-05 (Fisher's exact test), Q value = 0.0033

Table S155. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	47	122	439
subtype1	27	57	210
subtype2	14	59	132
subtype3	6	6	97

Figure S143. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.602 (Fisher's exact test), Q value = 1

Table S156. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	26	560
subtype1	13	268
subtype2	7	194
subtype3	6	98

Figure S144. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'ETHNICITY'

Methods & Data

Input

Cluster data file = BRCA-TP.mergedcluster.txt
Clinical data file = BRCA-TP.merged_data.txt
Number of patients = 1043
Number of clustering approaches = 12
Number of selected clinical features = 12
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