Correlation between aggregated molecular cancer subtypes and selected clinical features

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 13 clinical features across 274 patients, 28 significant findings detected with P value < 0.05 and Q value < 0.25.

3 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'PATHOLOGY.T.STAGE', 'PATHOLOGY.N.STAGE', 'COMPLETENESS.OF.RESECTION', 'NUMBER.OF.LYMPH.NODES', 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', and 'GLEASON_SCORE'.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', 'GLEASON_SCORE', and 'PSA_RESULT_PREOP'.
CNMF clustering analysis on RPPA data identified 3 subtypes that correlate to 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', and 'GLEASON_SCORE'.
Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that correlate to 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', and 'GLEASON_SCORE'.
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 'GLEASON_SCORE_PRIMARY'.
5 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'HISTOLOGICAL.TYPE' and 'GLEASON_SCORE_PRIMARY'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'PATHOLOGY.T.STAGE', 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', and 'GLEASON_SCORE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes do not correlate to any clinical features.
5 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'PATHOLOGY.T.STAGE', 'GLEASON_SCORE_COMBINED', 'GLEASON_SCORE_PRIMARY', and 'GLEASON_SCORE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 10 different clustering approaches and 13 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 28 significant findings detected.


Clinical Features	Statistical Tests	Copy Number Ratio CNMF subtypes	METHLYATION CNMF	RPPA CNMF subtypes	RPPA cHierClus subtypes	RNAseq CNMF subtypes	RNAseq cHierClus subtypes	MIRSEQ CNMF	MIRSEQ CHIERARCHICAL	MIRseq Mature CNMF subtypes	MIRseq Mature cHierClus subtypes
AGE	Kruskal-Wallis (anova)	0.215 (1.00)	0.00373 (0.366)	0.0791 (1.00)	0.0149 (1.00)	0.496 (1.00)	0.57 (1.00)	0.379 (1.00)	0.301 (1.00)	0.0573 (1.00)	0.636 (1.00)
PATHOLOGY T STAGE	Fisher's exact test	2e-05 (0.00244)	0.0486 (1.00)	0.0061 (0.573)	0.0204 (1.00)	0.017 (1.00)	0.0087 (0.783)	0.0124 (1.00)	0.00014 (0.0164)	0.419 (1.00)	8e-05 (0.00952)
PATHOLOGY N STAGE	Fisher's exact test	0.00016 (0.0184)	0.0138 (1.00)	0.0542 (1.00)	0.0163 (1.00)	0.00909 (0.809)	0.112 (1.00)	0.193 (1.00)	0.00299 (0.305)	0.113 (1.00)	0.0916 (1.00)
HISTOLOGICAL TYPE	Fisher's exact test	0.828 (1.00)	0.678 (1.00)	1 (1.00)	0.336 (1.00)	0.284 (1.00)	0.33 (1.00)	0.00064 (0.0691)	0.577 (1.00)	0.0408 (1.00)	0.014 (1.00)
COMPLETENESS OF RESECTION	Fisher's exact test	0.00112 (0.119)	0.461 (1.00)	0.0204 (1.00)	0.00354 (0.35)	0.438 (1.00)	0.251 (1.00)	0.0301 (1.00)	0.768 (1.00)	0.435 (1.00)	0.588 (1.00)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.000169 (0.0193)	0.0235 (1.00)	0.0381 (1.00)	0.0119 (1.00)	0.016 (1.00)	0.0887 (1.00)	0.148 (1.00)	0.00463 (0.444)	0.0953 (1.00)	0.134 (1.00)
GLEASON SCORE COMBINED	Kruskal-Wallis (anova)	1.39e-09 (1.8e-07)	0.00111 (0.119)	0.00011 (0.013)	4.87e-05 (0.00584)	0.0543 (1.00)	0.0062 (0.577)	0.0301 (1.00)	8.63e-06 (0.00109)	0.0034 (0.34)	0.000299 (0.0332)
GLEASON SCORE PRIMARY	Kruskal-Wallis (anova)	2.38e-09 (3.04e-07)	0.000468 (0.051)	1.56e-05 (0.00194)	1.14e-05 (0.00142)	0.00443 (0.43)	0.000183 (0.0207)	0.000358 (0.0394)	1.63e-05 (0.00201)	0.00568 (0.54)	0.00148 (0.154)
GLEASON SCORE SECONDARY	Kruskal-Wallis (anova)	0.0806 (1.00)	0.423 (1.00)	0.45 (1.00)	0.157 (1.00)	0.831 (1.00)	0.45 (1.00)	0.25 (1.00)	0.28 (1.00)	0.333 (1.00)	0.264 (1.00)
GLEASON SCORE	Kruskal-Wallis (anova)	2.08e-10 (2.71e-08)	0.0015 (0.154)	4.74e-05 (0.00573)	7.53e-06 (0.000956)	0.0301 (1.00)	0.0124 (1.00)	0.034 (1.00)	0.00014 (0.0164)	0.00761 (0.7)	0.00113 (0.119)
PSA RESULT PREOP	Kruskal-Wallis (anova)	0.0202 (1.00)	0.000198 (0.0222)	0.0443 (1.00)	0.0303 (1.00)	0.00335 (0.338)	0.02 (1.00)	0.0118 (1.00)	0.188 (1.00)	0.775 (1.00)	0.0185 (1.00)
PSA VALUE	Kruskal-Wallis (anova)	0.241 (1.00)	0.367 (1.00)	0.163 (1.00)	0.434 (1.00)	0.264 (1.00)	0.845 (1.00)	0.131 (1.00)	0.0193 (1.00)	0.668 (1.00)	0.149 (1.00)
RACE	Fisher's exact test	0.782 (1.00)	0.745 (1.00)	0.0271 (1.00)	0.00836 (0.761)	0.329 (1.00)	0.0104 (0.916)	0.417 (1.00)	0.824 (1.00)	0.828 (1.00)	0.139 (1.00)

Clustering Approach #1: 'Copy Number Ratio CNMF subtypes'

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

Cluster Labels	1	2	3
Number of samples	165	59	47

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.215 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	268	60.3 (7.0)
subtype1	162	59.8 (6.9)
subtype2	59	61.3 (6.5)
subtype3	47	60.7 (7.7)

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

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

P value = 2e-05 (Fisher's exact test), Q value = 0.0024

Table S3. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	128	136	5
subtype1	89	72	2
subtype2	12	45	2
subtype3	27	19	1

Figure S2. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.00016 (Fisher's exact test), Q value = 0.018

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

nPatients	0	1
ALL	201	24
subtype1	123	5
subtype2	43	14
subtype3	35	5

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

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

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

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

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	5	266
subtype1	4	161
subtype2	1	58
subtype3	0	47

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

'Copy Number Ratio CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.00112 (Fisher's exact test), Q value = 0.12

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

nPatients	R0	R1	R2	RX
ALL	192	57	3	6
subtype1	129	21	2	5
subtype2	35	20	0	0
subtype3	28	16	1	1

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

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

P value = 0.000169 (Kruskal-Wallis (anova)), Q value = 0.019

Table S7. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	223	0.2 (1.2)
subtype1	126	0.1 (0.4)
subtype2	57	0.7 (2.2)
subtype3	40	0.1 (0.3)

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

'Copy Number Ratio CNMF subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 1.39e-09 (Kruskal-Wallis (anova)), Q value = 1.8e-07

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

	nPatients	Mean (Std.Dev)
ALL	271	7.3 (0.8)
subtype1	165	7.1 (0.7)
subtype2	59	7.9 (0.9)
subtype3	47	7.2 (0.6)

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

'Copy Number Ratio CNMF subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 2.38e-09 (Kruskal-Wallis (anova)), Q value = 3e-07

Table S9. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	271	3.5 (0.6)
subtype1	165	3.4 (0.5)
subtype2	59	3.9 (0.6)
subtype3	47	3.4 (0.5)

Figure S8. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'Copy Number Ratio CNMF subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.0806 (Kruskal-Wallis (anova)), Q value = 1

Table S10. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	271	3.8 (0.6)
subtype1	165	3.7 (0.6)
subtype2	59	4.0 (0.7)
subtype3	47	3.8 (0.6)

Figure S9. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'Copy Number Ratio CNMF subtypes' versus 'GLEASON_SCORE'

P value = 2.08e-10 (Kruskal-Wallis (anova)), Q value = 2.7e-08

Table S11. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	271	7.3 (0.8)
subtype1	165	7.1 (0.7)
subtype2	59	8.0 (0.9)
subtype3	47	7.2 (0.7)

Figure S10. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'Copy Number Ratio CNMF subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.0202 (Kruskal-Wallis (anova)), Q value = 1

Table S12. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	269	10.4 (11.0)
subtype1	164	9.3 (10.7)
subtype2	59	14.1 (13.6)
subtype3	46	9.4 (6.8)

Figure S11. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'Copy Number Ratio CNMF subtypes' versus 'PSA_VALUE'

P value = 0.241 (Kruskal-Wallis (anova)), Q value = 1

Table S13. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	226	1.1 (3.9)
subtype1	144	0.8 (2.8)
subtype2	46	2.1 (6.5)
subtype3	36	1.0 (3.1)

Figure S12. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.782 (Fisher's exact test), Q value = 1

Table S14. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	146
subtype1	2	3	87
subtype2	0	2	32
subtype3	0	2	27

Figure S13. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #2: 'METHLYATION CNMF'

Table S15. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3
Number of samples	75	89	101

'METHLYATION CNMF' versus 'AGE'

P value = 0.00373 (Kruskal-Wallis (anova)), Q value = 0.37

Table S16. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	262	60.2 (7.0)
subtype1	74	61.9 (7.0)
subtype2	88	58.5 (7.0)
subtype3	100	60.5 (6.8)

Figure S14. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'AGE'

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

P value = 0.0486 (Fisher's exact test), Q value = 1

Table S17. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	123	135	5
subtype1	35	37	3
subtype2	49	38	0
subtype3	39	60	2

Figure S15. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.0138 (Fisher's exact test), Q value = 1

Table S18. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	194	24
subtype1	55	8
subtype2	68	2
subtype3	71	14

Figure S16. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'METHLYATION CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 0.678 (Fisher's exact test), Q value = 1

Table S19. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	6	259
subtype1	2	73
subtype2	1	88
subtype3	3	98

Figure S17. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'METHLYATION CNMF' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.461 (Fisher's exact test), Q value = 1

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	188	56	2	6
subtype1	49	21	0	1
subtype2	66	16	0	2
subtype3	73	19	2	3

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0235 (Kruskal-Wallis (anova)), Q value = 1

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	216	0.2 (1.2)
subtype1	63	0.2 (0.5)
subtype2	68	0.0 (0.2)
subtype3	85	0.5 (1.8)

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'METHLYATION CNMF' versus 'GLEASON_SCORE_COMBINED'

P value = 0.00111 (Kruskal-Wallis (anova)), Q value = 0.12

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	265	7.3 (0.8)
subtype1	75	7.5 (0.9)
subtype2	89	7.0 (0.6)
subtype3	101	7.4 (0.9)

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'METHLYATION CNMF' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.000468 (Kruskal-Wallis (anova)), Q value = 0.051

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	265	3.5 (0.6)
subtype1	75	3.7 (0.6)
subtype2	89	3.3 (0.5)
subtype3	101	3.5 (0.6)

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'METHLYATION CNMF' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.423 (Kruskal-Wallis (anova)), Q value = 1

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	265	3.8 (0.6)
subtype1	75	3.9 (0.7)
subtype2	89	3.7 (0.5)
subtype3	101	3.8 (0.6)

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'METHLYATION CNMF' versus 'GLEASON_SCORE'

P value = 0.0015 (Kruskal-Wallis (anova)), Q value = 0.15

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	265	7.3 (0.8)
subtype1	75	7.6 (1.0)
subtype2	89	7.1 (0.6)
subtype3	101	7.4 (0.9)

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'GLEASON_SCORE'

'METHLYATION CNMF' versus 'PSA_RESULT_PREOP'

P value = 0.000198 (Kruskal-Wallis (anova)), Q value = 0.022

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	263	10.3 (10.5)
subtype1	75	14.8 (15.4)
subtype2	89	6.9 (3.7)
subtype3	99	9.9 (8.7)

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'METHLYATION CNMF' versus 'PSA_VALUE'

P value = 0.367 (Kruskal-Wallis (anova)), Q value = 1

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	221	1.1 (3.9)
subtype1	57	1.3 (3.9)
subtype2	77	0.2 (1.2)
subtype3	87	1.7 (5.2)

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'PSA_VALUE'

'METHLYATION CNMF' versus 'RACE'

P value = 0.745 (Fisher's exact test), Q value = 1

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	147
subtype1	0	3	40
subtype2	1	1	48
subtype3	1	3	59

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

Clustering Approach #3: 'RPPA CNMF subtypes'

Table S29. Description of clustering approach #3: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	42	55	61

'RPPA CNMF subtypes' versus 'AGE'

P value = 0.0791 (Kruskal-Wallis (anova)), Q value = 1

Table S30. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	156	60.6 (7.1)
subtype1	42	62.7 (6.3)
subtype2	54	59.2 (7.8)
subtype3	60	60.4 (6.7)

Figure S27. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 0.0061 (Fisher's exact test), Q value = 0.57

Table S31. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	58	94	5
subtype1	18	23	0
subtype2	27	27	1
subtype3	13	44	4

Figure S28. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.0542 (Fisher's exact test), Q value = 1

Table S32. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	123	17
subtype1	38	4
subtype2	43	2
subtype3	42	11

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

'RPPA CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

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

Table S33. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	2	156
subtype1	0	42
subtype2	1	54
subtype3	1	60

Figure S30. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'RPPA CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0204 (Fisher's exact test), Q value = 1

Table S34. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	111	33	1	4
subtype1	36	4	0	1
subtype2	40	9	0	2
subtype3	35	20	1	1

Figure S31. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0381 (Kruskal-Wallis (anova)), Q value = 1

Table S35. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	140	0.2 (0.7)
subtype1	42	0.1 (0.5)
subtype2	45	0.1 (0.3)
subtype3	53	0.4 (1.1)

Figure S32. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'RPPA CNMF subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 0.00011 (Kruskal-Wallis (anova)), Q value = 0.013

Table S36. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	158	7.4 (0.9)
subtype1	42	7.1 (0.5)
subtype2	55	7.3 (0.8)
subtype3	61	7.8 (1.0)

Figure S33. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'RPPA CNMF subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 1.56e-05 (Kruskal-Wallis (anova)), Q value = 0.0019

Table S37. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	158	3.5 (0.6)
subtype1	42	3.3 (0.5)
subtype2	55	3.4 (0.5)
subtype3	61	3.8 (0.6)

Figure S34. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'RPPA CNMF subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.45 (Kruskal-Wallis (anova)), Q value = 1

Table S38. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	158	3.9 (0.7)
subtype1	42	3.7 (0.5)
subtype2	55	3.9 (0.7)
subtype3	61	3.9 (0.8)

Figure S35. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'RPPA CNMF subtypes' versus 'GLEASON_SCORE'

P value = 4.74e-05 (Kruskal-Wallis (anova)), Q value = 0.0057

Table S39. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	158	7.4 (0.9)
subtype1	42	7.1 (0.6)
subtype2	55	7.3 (0.8)
subtype3	61	7.9 (1.0)

Figure S36. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'RPPA CNMF subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.0443 (Kruskal-Wallis (anova)), Q value = 1

Table S40. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	156	11.3 (11.2)
subtype1	41	8.3 (5.0)
subtype2	54	10.5 (13.4)
subtype3	61	14.0 (11.7)

Figure S37. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'RPPA CNMF subtypes' versus 'PSA_VALUE'

P value = 0.163 (Kruskal-Wallis (anova)), Q value = 1

Table S41. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	143	1.4 (4.5)
subtype1	41	1.0 (2.8)
subtype2	49	0.8 (3.4)
subtype3	53	2.4 (6.0)

Figure S38. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.0271 (Fisher's exact test), Q value = 1

Table S42. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	6	121
subtype1	0	0	39
subtype2	2	1	42
subtype3	0	5	40

Figure S39. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S43. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	48	64	46

'RPPA cHierClus subtypes' versus 'AGE'

P value = 0.0149 (Kruskal-Wallis (anova)), Q value = 1

Table S44. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	156	60.6 (7.1)
subtype1	48	63.0 (6.2)
subtype2	63	60.5 (6.7)
subtype3	45	58.2 (7.8)

Figure S40. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 0.0204 (Fisher's exact test), Q value = 1

Table S45. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	58	94	5
subtype1	22	25	0
subtype2	15	45	4
subtype3	21	24	1

Figure S41. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.0163 (Fisher's exact test), Q value = 1

Table S46. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	123	17
subtype1	43	4
subtype2	43	12
subtype3	37	1

Figure S42. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.336 (Fisher's exact test), Q value = 1

Table S47. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	2	156
subtype1	0	48
subtype2	2	62
subtype3	0	46

Figure S43. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'RPPA cHierClus subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.00354 (Fisher's exact test), Q value = 0.35

Table S48. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	111	33	1	4
subtype1	43	4	0	0
subtype2	35	20	1	3
subtype3	33	9	0	1

Figure S44. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0119 (Kruskal-Wallis (anova)), Q value = 1

Table S49. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	140	0.2 (0.7)
subtype1	47	0.1 (0.5)
subtype2	55	0.4 (1.1)
subtype3	38	0.0 (0.2)

Figure S45. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'RPPA cHierClus subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 4.87e-05 (Kruskal-Wallis (anova)), Q value = 0.0058

Table S50. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	158	7.4 (0.9)
subtype1	48	7.0 (0.5)
subtype2	64	7.8 (1.0)
subtype3	46	7.3 (0.9)

Figure S46. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'RPPA cHierClus subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 1.14e-05 (Kruskal-Wallis (anova)), Q value = 0.0014

Table S51. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	158	3.5 (0.6)
subtype1	48	3.4 (0.5)
subtype2	64	3.8 (0.6)
subtype3	46	3.4 (0.5)

Figure S47. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'RPPA cHierClus subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.157 (Kruskal-Wallis (anova)), Q value = 1

Table S52. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	158	3.9 (0.7)
subtype1	48	3.7 (0.6)
subtype2	64	3.9 (0.8)
subtype3	46	3.9 (0.6)

Figure S48. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'RPPA cHierClus subtypes' versus 'GLEASON_SCORE'

P value = 7.53e-06 (Kruskal-Wallis (anova)), Q value = 0.00096

Table S53. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	158	7.4 (0.9)
subtype1	48	7.1 (0.6)
subtype2	64	7.9 (1.0)
subtype3	46	7.3 (0.9)

Figure S49. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'RPPA cHierClus subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.0303 (Kruskal-Wallis (anova)), Q value = 1

Table S54. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	156	11.3 (11.2)
subtype1	47	8.9 (6.4)
subtype2	64	14.1 (12.3)
subtype3	45	9.7 (12.9)

Figure S50. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'RPPA cHierClus subtypes' versus 'PSA_VALUE'

P value = 0.434 (Kruskal-Wallis (anova)), Q value = 1

Table S55. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	143	1.4 (4.5)
subtype1	47	1.4 (3.7)
subtype2	55	2.0 (5.7)
subtype3	41	0.6 (3.1)

Figure S51. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.00836 (Fisher's exact test), Q value = 0.76

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	6	121
subtype1	0	0	45
subtype2	0	5	40
subtype3	2	1	36

Figure S52. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #5: 'RNAseq CNMF subtypes'

Table S57. Description of clustering approach #5: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	86	85	97

'RNAseq CNMF subtypes' versus 'AGE'

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

Table S58. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	265	60.2 (7.0)
subtype1	84	60.7 (7.0)
subtype2	85	59.5 (7.1)
subtype3	96	60.4 (6.8)

Figure S53. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 0.017 (Fisher's exact test), Q value = 1

Table S59. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	134	5
subtype1	47	36	3
subtype2	45	37	1
subtype3	35	61	1

Figure S54. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.00909 (Fisher's exact test), Q value = 0.81

Table S60. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	198	24
subtype1	62	8
subtype2	69	2
subtype3	67	14

Figure S55. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

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

Table S61. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	5	263
subtype1	3	83
subtype2	0	85
subtype3	2	95

Figure S56. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'RNAseq CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.438 (Fisher's exact test), Q value = 1

Table S62. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	190	56	3	6
subtype1	57	24	1	1
subtype2	63	14	0	2
subtype3	70	18	2	3

Figure S57. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.016 (Kruskal-Wallis (anova)), Q value = 1

Table S63. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	220	0.2 (1.2)
subtype1	70	0.2 (0.8)
subtype2	69	0.0 (0.2)
subtype3	81	0.5 (1.8)

Figure S58. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'RNAseq CNMF subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 0.0543 (Kruskal-Wallis (anova)), Q value = 1

Table S64. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	268	7.3 (0.8)
subtype1	86	7.4 (0.9)
subtype2	85	7.1 (0.6)
subtype3	97	7.4 (0.9)

Figure S59. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'RNAseq CNMF subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.00443 (Kruskal-Wallis (anova)), Q value = 0.43

Table S65. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	268	3.5 (0.6)
subtype1	86	3.6 (0.6)
subtype2	85	3.3 (0.5)
subtype3	97	3.6 (0.6)

Figure S60. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'RNAseq CNMF subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.831 (Kruskal-Wallis (anova)), Q value = 1

Table S66. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	268	3.8 (0.6)
subtype1	86	3.8 (0.7)
subtype2	85	3.8 (0.5)
subtype3	97	3.8 (0.6)

Figure S61. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'RNAseq CNMF subtypes' versus 'GLEASON_SCORE'

P value = 0.0301 (Kruskal-Wallis (anova)), Q value = 1

Table S67. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	268	7.3 (0.8)
subtype1	86	7.4 (0.9)
subtype2	85	7.1 (0.6)
subtype3	97	7.4 (0.9)

Figure S62. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'RNAseq CNMF subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.00335 (Kruskal-Wallis (anova)), Q value = 0.34

Table S68. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	266	10.1 (10.2)
subtype1	86	13.3 (14.3)
subtype2	85	6.9 (3.5)
subtype3	95	10.0 (8.7)

Figure S63. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'RNAseq CNMF subtypes' versus 'PSA_VALUE'

P value = 0.264 (Kruskal-Wallis (anova)), Q value = 1

Table S69. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	224	1.1 (3.9)
subtype1	68	0.7 (3.0)
subtype2	74	0.6 (2.2)
subtype3	82	1.8 (5.4)

Figure S64. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.329 (Fisher's exact test), Q value = 1

Table S70. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	147
subtype1	0	3	38
subtype2	2	1	52
subtype3	0	3	57

Figure S65. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S71. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	97	82	89

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.57 (Kruskal-Wallis (anova)), Q value = 1

Table S72. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	265	60.2 (7.0)
subtype1	96	60.4 (7.2)
subtype2	81	59.5 (6.9)
subtype3	88	60.6 (6.8)

Figure S66. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 0.0087 (Fisher's exact test), Q value = 0.78

Table S73. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	134	5
subtype1	56	39	2
subtype2	40	40	0
subtype3	31	55	3

Figure S67. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.112 (Fisher's exact test), Q value = 1

Table S74. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	198	24
subtype1	71	8
subtype2	67	4
subtype3	60	12

Figure S68. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.33 (Fisher's exact test), Q value = 1

Table S75. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	5	263
subtype1	2	95
subtype2	0	82
subtype3	3	86

Figure S69. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'RNAseq cHierClus subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.251 (Fisher's exact test), Q value = 1

Table S76. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	190	56	3	6
subtype1	67	24	1	1
subtype2	64	11	0	2
subtype3	59	21	2	3

Figure S70. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0887 (Kruskal-Wallis (anova)), Q value = 1

Table S77. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	220	0.2 (1.2)
subtype1	79	0.2 (0.7)
subtype2	69	0.1 (0.2)
subtype3	72	0.5 (1.9)

Figure S71. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'RNAseq cHierClus subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 0.0062 (Kruskal-Wallis (anova)), Q value = 0.58

Table S78. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	268	7.3 (0.8)
subtype1	97	7.3 (0.8)
subtype2	82	7.1 (0.6)
subtype3	89	7.5 (0.9)

Figure S72. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'RNAseq cHierClus subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.000183 (Kruskal-Wallis (anova)), Q value = 0.021

Table S79. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	268	3.5 (0.6)
subtype1	97	3.5 (0.6)
subtype2	82	3.3 (0.5)
subtype3	89	3.7 (0.6)

Figure S73. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'RNAseq cHierClus subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.45 (Kruskal-Wallis (anova)), Q value = 1

Table S80. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	268	3.8 (0.6)
subtype1	97	3.7 (0.6)
subtype2	82	3.8 (0.5)
subtype3	89	3.9 (0.7)

Figure S74. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'RNAseq cHierClus subtypes' versus 'GLEASON_SCORE'

P value = 0.0124 (Kruskal-Wallis (anova)), Q value = 1

Table S81. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	268	7.3 (0.8)
subtype1	97	7.3 (0.8)
subtype2	82	7.1 (0.6)
subtype3	89	7.5 (1.0)

Figure S75. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'RNAseq cHierClus subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.02 (Kruskal-Wallis (anova)), Q value = 1

Table S82. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	266	10.1 (10.2)
subtype1	97	12.2 (12.4)
subtype2	82	7.5 (4.4)
subtype3	87	10.3 (10.9)

Figure S76. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'RNAseq cHierClus subtypes' versus 'PSA_VALUE'

P value = 0.845 (Kruskal-Wallis (anova)), Q value = 1

Table S83. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	224	1.1 (3.9)
subtype1	77	0.8 (3.0)
subtype2	74	0.8 (2.9)
subtype3	73	1.7 (5.3)

Figure S77. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.0104 (Fisher's exact test), Q value = 0.92

Table S84. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	147
subtype1	0	5	37
subtype2	2	0	63
subtype3	0	2	47

Figure S78. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S85. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3	4	5
Number of samples	50	23	58	60	81

'MIRSEQ CNMF' versus 'AGE'

P value = 0.379 (Kruskal-Wallis (anova)), Q value = 1

Table S86. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	269	60.4 (7.0)
subtype1	50	59.0 (7.2)
subtype2	23	60.0 (7.3)
subtype3	58	60.9 (7.1)
subtype4	60	61.6 (6.5)
subtype5	78	59.9 (7.1)

Figure S79. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'AGE'

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

P value = 0.0124 (Fisher's exact test), Q value = 1

Table S87. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	139	4
subtype1	23	25	0
subtype2	16	6	1
subtype3	22	36	0
subtype4	23	34	3
subtype5	43	38	0

Figure S80. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.193 (Fisher's exact test), Q value = 1

Table S88. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	203	23
subtype1	39	2
subtype2	17	1
subtype3	42	6
subtype4	46	10
subtype5	59	4

Figure S81. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'MIRSEQ CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 0.00064 (Fisher's exact test), Q value = 0.069

Table S89. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	6	266
subtype1	0	50
subtype2	0	23
subtype3	6	52
subtype4	0	60
subtype5	0	81

Figure S82. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRSEQ CNMF' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0301 (Fisher's exact test), Q value = 1

Table S90. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	193	57	3	6
subtype1	35	9	0	2
subtype2	15	6	1	0
subtype3	35	19	0	2
subtype4	45	15	0	0
subtype5	63	8	2	2

Figure S83. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.148 (Kruskal-Wallis (anova)), Q value = 1

Table S91. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	224	0.2 (1.2)
subtype1	39	0.1 (0.2)
subtype2	18	0.1 (0.2)
subtype3	48	0.5 (2.2)
subtype4	56	0.4 (1.1)
subtype5	63	0.1 (0.3)

Figure S84. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CNMF' versus 'GLEASON_SCORE_COMBINED'

P value = 0.0301 (Kruskal-Wallis (anova)), Q value = 1

Table S92. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	50	7.1 (0.6)
subtype2	23	7.2 (1.0)
subtype3	58	7.5 (1.0)
subtype4	60	7.4 (0.8)
subtype5	81	7.1 (0.7)

Figure S85. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'MIRSEQ CNMF' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.000358 (Kruskal-Wallis (anova)), Q value = 0.039

Table S93. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.5 (0.6)
subtype1	50	3.3 (0.5)
subtype2	23	3.3 (0.5)
subtype3	58	3.6 (0.6)
subtype4	60	3.7 (0.6)
subtype5	81	3.4 (0.5)

Figure S86. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'MIRSEQ CNMF' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.25 (Kruskal-Wallis (anova)), Q value = 1

Table S94. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.8 (0.6)
subtype1	50	3.8 (0.5)
subtype2	23	3.9 (0.6)
subtype3	58	3.9 (0.7)
subtype4	60	3.8 (0.7)
subtype5	81	3.7 (0.6)

Figure S87. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'MIRSEQ CNMF' versus 'GLEASON_SCORE'

P value = 0.034 (Kruskal-Wallis (anova)), Q value = 1

Table S95. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	50	7.2 (0.6)
subtype2	23	7.2 (1.1)
subtype3	58	7.6 (1.0)
subtype4	60	7.5 (0.8)
subtype5	81	7.2 (0.7)

Figure S88. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'GLEASON_SCORE'

'MIRSEQ CNMF' versus 'PSA_RESULT_PREOP'

P value = 0.0118 (Kruskal-Wallis (anova)), Q value = 1

Table S96. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	270	10.5 (11.2)
subtype1	50	7.2 (3.7)
subtype2	23	11.6 (16.3)
subtype3	58	11.7 (14.0)
subtype4	58	13.9 (14.1)
subtype5	81	8.8 (6.3)

Figure S89. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'MIRSEQ CNMF' versus 'PSA_VALUE'

P value = 0.131 (Kruskal-Wallis (anova)), Q value = 1

Table S97. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	228	1.1 (3.9)
subtype1	44	0.7 (2.6)
subtype2	22	0.4 (1.7)
subtype3	42	1.1 (4.2)
subtype4	54	2.4 (6.1)
subtype5	66	0.5 (1.8)

Figure S90. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'PSA_VALUE'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.417 (Fisher's exact test), Q value = 1

Table S98. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	146
subtype1	0	0	25
subtype2	0	1	8
subtype3	0	0	2
subtype4	1	1	50
subtype5	1	5	61

Figure S91. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S99. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	94	50	128

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.301 (Kruskal-Wallis (anova)), Q value = 1

Table S100. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	269	60.4 (7.0)
subtype1	94	59.7 (6.5)
subtype2	49	60.0 (7.2)
subtype3	126	61.0 (7.3)

Figure S92. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'AGE'

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

P value = 0.00014 (Fisher's exact test), Q value = 0.016

Table S101. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	139	4
subtype1	44	48	0
subtype2	31	15	4
subtype3	52	76	0

Figure S93. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.00299 (Fisher's exact test), Q value = 0.3

Table S102. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	203	23
subtype1	77	5
subtype2	37	0
subtype3	89	18

Figure S94. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL.TYPE'

P value = 0.577 (Fisher's exact test), Q value = 1

Table S103. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	6	266
subtype1	3	91
subtype2	0	50
subtype3	3	125

Figure S95. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRSEQ CHIERARCHICAL' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.768 (Fisher's exact test), Q value = 1

Table S104. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	193	57	3	6
subtype1	70	16	0	2
subtype2	34	12	1	1
subtype3	89	29	2	3

Figure S96. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.00463 (Kruskal-Wallis (anova)), Q value = 0.44

Table S105. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	224	0.2 (1.2)
subtype1	80	0.1 (0.4)
subtype2	37	0.0 (0.0)
subtype3	107	0.4 (1.6)

Figure S97. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CHIERARCHICAL' versus 'GLEASON_SCORE_COMBINED'

P value = 8.63e-06 (Kruskal-Wallis (anova)), Q value = 0.0011

Table S106. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	94	7.1 (0.7)
subtype2	50	7.0 (0.7)
subtype3	128	7.5 (0.9)

Figure S98. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'MIRSEQ CHIERARCHICAL' versus 'GLEASON_SCORE_PRIMARY'

P value = 1.63e-05 (Kruskal-Wallis (anova)), Q value = 0.002

Table S107. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.5 (0.6)
subtype1	94	3.4 (0.5)
subtype2	50	3.3 (0.5)
subtype3	128	3.7 (0.6)

Figure S99. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'MIRSEQ CHIERARCHICAL' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.28 (Kruskal-Wallis (anova)), Q value = 1

Table S108. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.8 (0.6)
subtype1	94	3.8 (0.6)
subtype2	50	3.7 (0.6)
subtype3	128	3.9 (0.7)

Figure S100. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'MIRSEQ CHIERARCHICAL' versus 'GLEASON_SCORE'

P value = 0.00014 (Kruskal-Wallis (anova)), Q value = 0.016

Table S109. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	94	7.1 (0.7)
subtype2	50	7.1 (0.8)
subtype3	128	7.5 (0.9)

Figure S101. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'GLEASON_SCORE'

'MIRSEQ CHIERARCHICAL' versus 'PSA_RESULT_PREOP'

P value = 0.188 (Kruskal-Wallis (anova)), Q value = 1

Table S110. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	270	10.5 (11.2)
subtype1	93	9.1 (10.3)
subtype2	50	10.6 (13.2)
subtype3	127	11.4 (11.0)

Figure S102. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'MIRSEQ CHIERARCHICAL' versus 'PSA_VALUE'

P value = 0.0193 (Kruskal-Wallis (anova)), Q value = 1

Table S111. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	228	1.1 (3.9)
subtype1	85	1.0 (3.0)
subtype2	47	0.1 (0.2)
subtype3	96	1.7 (5.2)

Figure S103. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'PSA_VALUE'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.824 (Fisher's exact test), Q value = 1

Table S112. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	146
subtype1	1	1	50
subtype2	0	2	28
subtype3	1	4	68

Figure S104. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S113. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	73	99	100

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.0573 (Kruskal-Wallis (anova)), Q value = 1

Table S114. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	269	60.4 (7.0)
subtype1	73	62.0 (6.7)
subtype2	96	59.3 (7.1)
subtype3	100	60.2 (7.0)

Figure S105. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 0.419 (Fisher's exact test), Q value = 1

Table S115. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	139	4
subtype1	31	42	0
subtype2	51	47	1
subtype3	45	50	3

Figure S106. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.113 (Fisher's exact test), Q value = 1

Table S116. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	203	23
subtype1	50	10
subtype2	72	8
subtype3	81	5

Figure S107. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

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

P value = 0.0408 (Fisher's exact test), Q value = 1

Table S117. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	6	266
subtype1	4	69
subtype2	0	99
subtype3	2	98

Figure S108. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRseq Mature CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.435 (Fisher's exact test), Q value = 1

Table S118. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	193	57	3	6
subtype1	48	21	0	2
subtype2	73	15	2	2
subtype3	72	21	1	2

Figure S109. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0953 (Kruskal-Wallis (anova)), Q value = 1

Table S119. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	224	0.2 (1.2)
subtype1	60	0.4 (1.0)
subtype2	80	0.3 (1.7)
subtype3	84	0.1 (0.4)

Figure S110. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature CNMF subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 0.0034 (Kruskal-Wallis (anova)), Q value = 0.34

Table S120. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	73	7.6 (0.9)
subtype2	99	7.2 (0.7)
subtype3	100	7.2 (0.7)

Figure S111. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'MIRseq Mature CNMF subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.00568 (Kruskal-Wallis (anova)), Q value = 0.54

Table S121. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.5 (0.6)
subtype1	73	3.7 (0.6)
subtype2	99	3.4 (0.5)
subtype3	100	3.4 (0.5)

Figure S112. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'MIRseq Mature CNMF subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.333 (Kruskal-Wallis (anova)), Q value = 1

Table S122. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.8 (0.6)
subtype1	73	3.9 (0.7)
subtype2	99	3.8 (0.6)
subtype3	100	3.8 (0.6)

Figure S113. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'MIRseq Mature CNMF subtypes' versus 'GLEASON_SCORE'

P value = 0.00761 (Kruskal-Wallis (anova)), Q value = 0.7

Table S123. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	73	7.6 (1.0)
subtype2	99	7.2 (0.7)
subtype3	100	7.2 (0.8)

Figure S114. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'MIRseq Mature CNMF subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.775 (Kruskal-Wallis (anova)), Q value = 1

Table S124. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	270	10.5 (11.2)
subtype1	72	12.0 (12.4)
subtype2	99	10.0 (10.7)
subtype3	99	9.9 (10.8)

Figure S115. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'MIRseq Mature CNMF subtypes' versus 'PSA_VALUE'

P value = 0.668 (Kruskal-Wallis (anova)), Q value = 1

Table S125. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	228	1.1 (3.9)
subtype1	54	1.7 (4.4)
subtype2	81	0.5 (1.6)
subtype3	93	1.3 (4.8)

Figure S116. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'MIRseq Mature CNMF subtypes' versus 'RACE'

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

Table S126. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	146
subtype1	0	0	19
subtype2	1	5	69
subtype3	1	2	58

Figure S117. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S127. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	44	25	55	62	86

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.636 (Kruskal-Wallis (anova)), Q value = 1

Table S128. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	269	60.4 (7.0)
subtype1	44	60.6 (6.6)
subtype2	25	59.4 (7.8)
subtype3	55	61.4 (7.0)
subtype4	62	60.7 (7.0)
subtype5	83	59.7 (7.1)

Figure S118. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'AGE'

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

P value = 8e-05 (Fisher's exact test), Q value = 0.0095

Table S129. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

nPatients	T2	T3	T4
ALL	127	139	4
subtype1	25	18	0
subtype2	14	7	4
subtype3	20	35	0
subtype4	22	39	0
subtype5	46	40	0

Figure S119. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'PATHOLOGY.T.STAGE'

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

P value = 0.0916 (Fisher's exact test), Q value = 1

Table S130. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	203	23
subtype1	33	0
subtype2	19	1
subtype3	40	7
subtype4	50	9
subtype5	61	6

Figure S120. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.N.STAGE'

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

P value = 0.014 (Fisher's exact test), Q value = 1

Table S131. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

nPatients	PROSTATE ADENOCARCINOMA OTHER SUBTYPE	PROSTATE ADENOCARCINOMA ACINAR TYPE
ALL	6	266
subtype1	3	41
subtype2	0	25
subtype3	3	52
subtype4	0	62
subtype5	0	86

Figure S121. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRseq Mature cHierClus subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.588 (Fisher's exact test), Q value = 1

Table S132. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	193	57	3	6
subtype1	29	10	0	1
subtype2	16	7	1	0
subtype3	37	15	0	2
subtype4	48	13	0	1
subtype5	63	12	2	2

Figure S122. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.134 (Kruskal-Wallis (anova)), Q value = 1

Table S133. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	224	0.2 (1.2)
subtype1	31	0.0 (0.0)
subtype2	20	0.1 (0.2)
subtype3	47	0.3 (1.0)
subtype4	59	0.3 (0.7)
subtype5	67	0.3 (1.9)

Figure S123. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature cHierClus subtypes' versus 'GLEASON_SCORE_COMBINED'

P value = 0.000299 (Kruskal-Wallis (anova)), Q value = 0.033

Table S134. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	44	7.1 (0.6)
subtype2	25	7.2 (0.9)
subtype3	55	7.7 (1.0)
subtype4	62	7.3 (0.8)
subtype5	86	7.2 (0.7)

Figure S124. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GLEASON_SCORE_COMBINED'

'MIRseq Mature cHierClus subtypes' versus 'GLEASON_SCORE_PRIMARY'

P value = 0.00148 (Kruskal-Wallis (anova)), Q value = 0.15

Table S135. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.5 (0.6)
subtype1	44	3.4 (0.5)
subtype2	25	3.5 (0.6)
subtype3	55	3.8 (0.6)
subtype4	62	3.4 (0.6)
subtype5	86	3.4 (0.5)

Figure S125. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'GLEASON_SCORE_PRIMARY'

'MIRseq Mature cHierClus subtypes' versus 'GLEASON_SCORE_SECONDARY'

P value = 0.264 (Kruskal-Wallis (anova)), Q value = 1

Table S136. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

	nPatients	Mean (Std.Dev)
ALL	272	3.8 (0.6)
subtype1	44	3.7 (0.5)
subtype2	25	3.7 (0.6)
subtype3	55	4.0 (0.7)
subtype4	62	3.8 (0.6)
subtype5	86	3.7 (0.6)

Figure S126. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'GLEASON_SCORE_SECONDARY'

'MIRseq Mature cHierClus subtypes' versus 'GLEASON_SCORE'

P value = 0.00113 (Kruskal-Wallis (anova)), Q value = 0.12

Table S137. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

	nPatients	Mean (Std.Dev)
ALL	272	7.3 (0.8)
subtype1	44	7.1 (0.6)
subtype2	25	7.4 (1.0)
subtype3	55	7.7 (1.0)
subtype4	62	7.3 (0.8)
subtype5	86	7.2 (0.7)

Figure S127. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'GLEASON_SCORE'

'MIRseq Mature cHierClus subtypes' versus 'PSA_RESULT_PREOP'

P value = 0.0185 (Kruskal-Wallis (anova)), Q value = 1

Table S138. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

	nPatients	Mean (Std.Dev)
ALL	270	10.5 (11.2)
subtype1	44	9.1 (13.8)
subtype2	25	13.3 (17.2)
subtype3	55	12.6 (12.8)
subtype4	60	10.8 (8.3)
subtype5	86	8.9 (7.5)

Figure S128. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'PSA_RESULT_PREOP'

'MIRseq Mature cHierClus subtypes' versus 'PSA_VALUE'

P value = 0.149 (Kruskal-Wallis (anova)), Q value = 1

Table S139. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

	nPatients	Mean (Std.Dev)
ALL	228	1.1 (3.9)
subtype1	37	0.6 (2.1)
subtype2	25	0.1 (0.2)
subtype3	36	1.7 (4.7)
subtype4	60	2.1 (6.0)
subtype5	70	0.5 (1.7)

Figure S129. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'PSA_VALUE'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.139 (Fisher's exact test), Q value = 1

Table S140. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	7	146
subtype1	0	0	4
subtype2	0	2	11
subtype3	0	1	9
subtype4	1	0	61
subtype5	1	4	61

Figure S130. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

Methods & Data

Input

Cluster data file = PRAD-TP.mergedcluster.txt
Clinical data file = PRAD-TP.merged_data.txt
Number of patients = 274
Number of clustering approaches = 10
Number of selected clinical features = 13
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)

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.

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] 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)

[4] 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