Correlation between aggregated molecular cancer subtypes and selected clinical features

Colorectal Adenocarcinoma (Primary solid tumor)

02 April 2015 | analyses__2015_04_02

Maintainer Information

Citation Information

Maintained by TCGA GDAC Team (Broad Institute/MD Anderson Cancer Center/Harvard Medical School)

Cite as Broad Institute TCGA Genome Data Analysis Center (2015): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C100013T

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 13 clinical features across 621 patients, 62 significant findings detected with P value < 0.05 and Q value < 0.25.

CNMF clustering analysis on array-based mRNA expression data identified 4 subtypes that correlate to 'PRIMARY_SITE_OF_DISEASE', 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_N_STAGE', 'PATHOLOGY_M_STAGE', 'HISTOLOGICAL_TYPE', 'COMPLETENESS_OF_RESECTION', and 'NUMBER_OF_LYMPH_NODES'.
Consensus hierarchical clustering analysis on array-based mRNA expression data identified 4 subtypes that correlate to 'YEARS_TO_BIRTH', 'PRIMARY_SITE_OF_DISEASE', 'NEOPLASM_DISEASESTAGE', 'HISTOLOGICAL_TYPE', and 'COMPLETENESS_OF_RESECTION'.
3 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'PRIMARY_SITE_OF_DISEASE', 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_N_STAGE', 'PATHOLOGY_M_STAGE', 'HISTOLOGICAL_TYPE', 'COMPLETENESS_OF_RESECTION', and 'NUMBER_OF_LYMPH_NODES'.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'YEARS_TO_BIRTH', 'PRIMARY_SITE_OF_DISEASE', 'GENDER', and 'HISTOLOGICAL_TYPE'.
CNMF clustering analysis on RPPA data identified 4 subtypes that correlate to 'PATHOLOGY_M_STAGE', 'COMPLETENESS_OF_RESECTION', 'NUMBER_OF_LYMPH_NODES', and 'RACE'.
Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that correlate to 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'PATHOLOGY_M_STAGE', 'COMPLETENESS_OF_RESECTION', and 'NUMBER_OF_LYMPH_NODES'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that correlate to 'YEARS_TO_BIRTH', 'NEOPLASM_DISEASESTAGE', 'HISTOLOGICAL_TYPE', and 'COMPLETENESS_OF_RESECTION'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 7 subtypes that correlate to 'YEARS_TO_BIRTH', 'PRIMARY_SITE_OF_DISEASE', 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'RADIATIONS_RADIATION_REGIMENINDICATION', 'COMPLETENESS_OF_RESECTION', and 'NUMBER_OF_LYMPH_NODES'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'Time to Death', 'YEARS_TO_BIRTH', 'PRIMARY_SITE_OF_DISEASE', 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'COMPLETENESS_OF_RESECTION', and 'NUMBER_OF_LYMPH_NODES'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'YEARS_TO_BIRTH', 'NEOPLASM_DISEASESTAGE', and 'COMPLETENESS_OF_RESECTION'.
6 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'PRIMARY_SITE_OF_DISEASE', 'PATHOLOGY_M_STAGE', 'HISTOLOGICAL_TYPE', and 'NUMBER_OF_LYMPH_NODES'.
4 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'PRIMARY_SITE_OF_DISEASE' and 'HISTOLOGICAL_TYPE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 12 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, 62 significant findings detected.


Clinical Features	Statistical Tests	mRNA CNMF subtypes	mRNA cHierClus subtypes	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
Time to Death	logrank test	0.326 (0.489)	0.233 (0.379)	0.53 (0.667)	0.938 (0.944)	0.432 (0.579)	0.26 (0.415)	0.262 (0.415)	0.533 (0.667)	0.0468 (0.123)	0.843 (0.871)	0.72 (0.81)	0.565 (0.687)
YEARS TO BIRTH	Kruskal-Wallis (anova)	0.434 (0.579)	0.0483 (0.123)	0.412 (0.559)	0.00177 (0.0102)	0.955 (0.955)	0.65 (0.762)	0.00149 (0.00928)	0.000798 (0.00566)	0.00875 (0.0414)	0.00428 (0.0215)	0.201 (0.345)	0.516 (0.663)
PRIMARY SITE OF DISEASE	Fisher's exact test	0.0246 (0.0818)	0.00302 (0.0162)	3e-05 (0.000312)	0.00098 (0.00637)	0.826 (0.864)	0.768 (0.832)	0.0866 (0.185)	1e-05 (0.000142)	0.0403 (0.113)	0.263 (0.415)	1e-05 (0.000142)	1e-05 (0.000142)
NEOPLASM DISEASESTAGE	Fisher's exact test	0.00578 (0.0282)	0.048 (0.123)	0.00419 (0.0215)	0.679 (0.779)	0.0995 (0.206)	0.0296 (0.0906)	0.00093 (0.00631)	1e-05 (0.000142)	0.00021 (0.00172)	0.00974 (0.0447)	0.196 (0.34)	0.767 (0.832)
PATHOLOGY T STAGE	Fisher's exact test	0.0565 (0.136)	0.0629 (0.142)	0.619 (0.737)	0.568 (0.687)	0.337 (0.499)	0.0426 (0.117)	0.0514 (0.125)	0.613 (0.736)	0.104 (0.212)	0.92 (0.932)	0.144 (0.271)	0.44 (0.582)
PATHOLOGY N STAGE	Fisher's exact test	0.0158 (0.0591)	0.461 (0.604)	0.00019 (0.00165)	0.225 (0.37)	0.0727 (0.16)	0.017 (0.0591)	0.155 (0.287)	0.0138 (0.0551)	0.0169 (0.0591)	0.144 (0.271)	0.217 (0.363)	0.751 (0.831)
PATHOLOGY M STAGE	Fisher's exact test	0.0264 (0.0823)	0.0848 (0.184)	0.00238 (0.0133)	0.402 (0.55)	0.0117 (0.0508)	0.0307 (0.0922)	0.108 (0.216)	0.0511 (0.125)	0.816 (0.861)	0.192 (0.337)	0.0257 (0.0819)	0.471 (0.612)
GENDER	Fisher's exact test	0.0603 (0.138)	0.187 (0.332)	0.281 (0.439)	0.0361 (0.104)	0.716 (0.81)	0.325 (0.489)	0.654 (0.762)	0.346 (0.503)	0.541 (0.67)	0.519 (0.663)	0.856 (0.879)	0.323 (0.489)
HISTOLOGICAL TYPE	Fisher's exact test	2e-05 (0.000223)	1e-05 (0.000142)	1e-05 (0.000142)	0.00011 (0.00101)	0.182 (0.326)	0.207 (0.351)	2e-05 (0.000223)	1e-05 (0.000142)	0.00031 (0.0023)	0.339 (0.499)	1e-05 (0.000142)	1e-05 (0.000142)
RADIATIONS RADIATION REGIMENINDICATION	Fisher's exact test	0.397 (0.55)	0.654 (0.762)	0.838 (0.871)	0.0592 (0.138)	0.731 (0.814)	0.0578 (0.137)	0.376 (0.533)	0.0407 (0.113)	0.0948 (0.2)	0.4 (0.55)	0.879 (0.896)	0.144 (0.271)
COMPLETENESS OF RESECTION	Fisher's exact test	0.0195 (0.0663)	0.0103 (0.0457)	0.0165 (0.0591)	0.319 (0.489)	0.0171 (0.0591)	0.0147 (0.0572)	2e-05 (0.000223)	1e-05 (0.000142)	1e-05 (0.000142)	8e-05 (0.00078)	0.535 (0.667)	0.356 (0.509)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.0127 (0.0537)	0.174 (0.316)	0.00164 (0.00987)	0.562 (0.687)	0.0256 (0.0819)	0.000274 (0.00214)	0.068 (0.151)	0.0134 (0.055)	0.0488 (0.123)	0.127 (0.251)	0.0318 (0.0935)	0.348 (0.503)
RACE	Fisher's exact test	0.135 (0.263)	0.167 (0.306)	0.722 (0.81)	0.382 (0.536)	0.0461 (0.123)	0.76 (0.832)	0.782 (0.842)	0.1 (0.206)	0.661 (0.764)	0.79 (0.844)	0.219 (0.363)	0.796 (0.844)

Clustering Approach #1: 'mRNA CNMF subtypes'

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

Cluster Labels	1	2	3	4
Number of samples	46	62	72	42

'mRNA CNMF subtypes' versus 'Time to Death'

P value = 0.326 (logrank test), Q value = 0.49

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	208	40	0.0 - 54.0 (21.0)
subtype1	41	9	0.0 - 46.6 (22.0)
subtype2	58	15	1.0 - 52.0 (20.5)
subtype3	68	10	0.0 - 50.0 (20.5)
subtype4	41	6	0.0 - 54.0 (23.0)

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

'mRNA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.434 (Kruskal-Wallis (anova)), Q value = 0.58

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

	nPatients	Mean (Std.Dev)
ALL	222	69.5 (11.4)
subtype1	46	70.8 (11.8)
subtype2	62	68.5 (8.9)
subtype3	72	69.5 (12.4)
subtype4	42	69.6 (12.6)

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

'mRNA CNMF subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.0246 (Fisher's exact test), Q value = 0.082

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

nPatients	COLON	RECTUM
ALL	152	68
subtype1	39	7
subtype2	37	25
subtype3	45	25
subtype4	31	11

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

'mRNA CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00578 (Fisher's exact test), Q value = 0.028

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

nPatients	STAGE I	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	47	15	65	5	10	3	22	20	33	1
subtype1	8	3	15	3	3	1	4	2	6	0
subtype2	9	2	14	0	2	2	5	11	17	0
subtype3	19	4	23	0	4	0	7	7	8	0
subtype4	11	6	13	2	1	0	6	0	2	1

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

'mRNA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0565 (Fisher's exact test), Q value = 0.14

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

nPatients	T1	T2	T3	T4
ALL	9	46	148	19
subtype1	1	9	30	6
subtype2	1	12	41	8
subtype3	5	16	51	0
subtype4	2	9	26	5

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

'mRNA CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0158 (Fisher's exact test), Q value = 0.059

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

nPatients	N0	N1	N2
ALL	136	43	43
subtype1	30	6	10
subtype2	28	17	17
subtype3	46	12	14
subtype4	32	8	2

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

'mRNA CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0264 (Fisher's exact test), Q value = 0.082

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

nPatients	0	1
ALL	186	34
subtype1	39	6
subtype2	45	17
subtype3	63	8
subtype4	39	3

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

'mRNA CNMF subtypes' versus 'GENDER'

P value = 0.0603 (Fisher's exact test), Q value = 0.14

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

nPatients	FEMALE	MALE
ALL	106	116
subtype1	28	18
subtype2	33	29
subtype3	30	42
subtype4	15	27

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

'mRNA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 2e-05 (Fisher's exact test), Q value = 0.00022

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

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	129	22	58	7
subtype1	28	11	4	2
subtype2	36	1	25	0
subtype3	44	1	21	2
subtype4	21	9	8	3

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

'mRNA CNMF subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.397 (Fisher's exact test), Q value = 0.55

Table S11. Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	1	221
subtype1	1	45
subtype2	0	62
subtype3	0	72
subtype4	0	42

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

'mRNA CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0195 (Fisher's exact test), Q value = 0.066

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

nPatients	R0	R1	R2
ALL	185	2	29
subtype1	38	0	4
subtype2	45	2	15
subtype3	62	0	8
subtype4	40	0	2

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

'mRNA CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0127 (Kruskal-Wallis (anova)), Q value = 0.054

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

	nPatients	Mean (Std.Dev)
ALL	221	2.2 (4.7)
subtype1	46	3.3 (7.0)
subtype2	62	2.9 (5.4)
subtype3	72	1.6 (2.7)
subtype4	41	0.8 (2.2)

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

'mRNA CNMF subtypes' versus 'RACE'

P value = 0.135 (Fisher's exact test), Q value = 0.26

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

nPatients	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	15
subtype1	0	7
subtype2	2	3
subtype3	0	4
subtype4	0	1

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

Clustering Approach #2: 'mRNA cHierClus subtypes'

Table S15. Description of clustering approach #2: 'mRNA cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	52	41	77	52

'mRNA cHierClus subtypes' versus 'Time to Death'

P value = 0.233 (logrank test), Q value = 0.38

Table S16. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	208	40	0.0 - 54.0 (21.0)
subtype1	48	11	0.0 - 53.0 (20.6)
subtype2	39	11	1.0 - 48.0 (24.0)
subtype3	74	13	0.0 - 54.0 (20.5)
subtype4	47	5	0.9 - 50.0 (22.0)

Figure S14. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'mRNA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0483 (Kruskal-Wallis (anova)), Q value = 0.12

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

	nPatients	Mean (Std.Dev)
ALL	222	69.5 (11.4)
subtype1	52	72.9 (11.6)
subtype2	41	67.7 (9.6)
subtype3	77	68.1 (11.7)
subtype4	52	69.5 (11.6)

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

'mRNA cHierClus subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.00302 (Fisher's exact test), Q value = 0.016

Table S18. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	152	68
subtype1	46	6
subtype2	24	17
subtype3	48	28
subtype4	34	17

Figure S16. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'mRNA cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.048 (Fisher's exact test), Q value = 0.12

Table S19. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	47	15	65	5	10	3	22	20	33	1
subtype1	9	3	17	4	3	1	6	2	5	1
subtype2	4	2	10	1	0	1	4	7	12	0
subtype3	20	8	23	0	6	1	5	5	9	0
subtype4	14	2	15	0	1	0	7	6	7	0

Figure S17. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0629 (Fisher's exact test), Q value = 0.14

Table S20. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	9	46	148	19
subtype1	1	10	33	8
subtype2	0	7	28	6
subtype3	7	15	51	4
subtype4	1	14	36	1

Figure S18. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.461 (Fisher's exact test), Q value = 0.6

Table S21. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	136	43	43
subtype1	34	9	9
subtype2	19	11	11
subtype3	52	12	13
subtype4	31	11	10

Figure S19. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'mRNA cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0848 (Fisher's exact test), Q value = 0.18

Table S22. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	186	34
subtype1	45	6
subtype2	29	12
subtype3	68	9
subtype4	44	7

Figure S20. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'mRNA cHierClus subtypes' versus 'GENDER'

P value = 0.187 (Fisher's exact test), Q value = 0.33

Table S23. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	106	116
subtype1	30	22
subtype2	22	19
subtype3	34	43
subtype4	20	32

Figure S21. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

'mRNA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S24. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	129	22	58	7
subtype1	27	18	2	3
subtype2	24	0	17	0
subtype3	44	4	23	3
subtype4	34	0	16	1

Figure S22. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'mRNA cHierClus subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.654 (Fisher's exact test), Q value = 0.76

Table S25. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	1	221
subtype1	1	51
subtype2	0	41
subtype3	0	77
subtype4	0	52

Figure S23. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'mRNA cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0103 (Fisher's exact test), Q value = 0.046

Table S26. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2
ALL	185	2	29
subtype1	45	0	3
subtype2	28	2	11
subtype3	69	0	8
subtype4	43	0	7

Figure S24. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'mRNA cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.174 (Kruskal-Wallis (anova)), Q value = 0.32

Table S27. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	221	2.2 (4.7)
subtype1	52	2.8 (6.4)
subtype2	41	3.3 (6.4)
subtype3	76	1.5 (2.9)
subtype4	52	1.6 (2.6)

Figure S25. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'mRNA cHierClus subtypes' versus 'RACE'

P value = 0.167 (Fisher's exact test), Q value = 0.31

Table S28. Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	15
subtype1	0	6
subtype2	2	3
subtype3	0	5
subtype4	0	1

Figure S26. Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

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

Table S29. Description of clustering approach #3: 'Copy Number Ratio CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	252	205	151

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

P value = 0.53 (logrank test), Q value = 0.67

Table S30. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	583	115	0.0 - 140.4 (19.1)
subtype1	243	53	0.0 - 140.4 (19.0)
subtype2	195	32	0.1 - 131.5 (19.0)
subtype3	145	30	0.0 - 130.7 (21.0)

Figure S27. 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 'YEARS_TO_BIRTH'

P value = 0.412 (Kruskal-Wallis (anova)), Q value = 0.56

Table S31. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	606	66.3 (12.7)
subtype1	250	66.8 (14.1)
subtype2	205	66.1 (11.2)
subtype3	151	65.7 (12.4)

Figure S28. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'Copy Number Ratio CNMF subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 3e-05 (Fisher's exact test), Q value = 0.00031

Table S32. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	442	162
subtype1	209	43
subtype2	130	71
subtype3	103	48

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

'Copy Number Ratio CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00419 (Fisher's exact test), Q value = 0.022

Table S33. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	104	1	36	171	10	3	26	21	79	54	61	27	2
subtype1	53	1	18	82	6	0	11	7	26	19	15	7	1
subtype2	33	0	5	54	3	1	9	7	34	14	26	13	1
subtype3	18	0	13	35	1	2	6	7	19	21	20	7	0

Figure S30. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.619 (Fisher's exact test), Q value = 0.74

Table S34. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	20	107	413	66
subtype1	10	50	163	28
subtype2	8	33	141	22
subtype3	2	24	109	16

Figure S31. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.00019 (Fisher's exact test), Q value = 0.0016

Table S35. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	342	149	114
subtype1	166	48	38
subtype2	103	66	34
subtype3	73	35	42

Figure S32. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.00238 (Fisher's exact test), Q value = 0.013

Table S36. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	448	86
subtype1	197	21
subtype2	144	39
subtype3	107	26

Figure S33. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.281 (Fisher's exact test), Q value = 0.44

Table S37. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	286	322
subtype1	127	125
subtype2	88	117
subtype3	71	80

Figure S34. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'GENDER'

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S38. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	381	60	146	13
subtype1	161	47	33	9
subtype2	125	5	71	0
subtype3	95	8	42	4

Figure S35. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.838 (Fisher's exact test), Q value = 0.87

Table S39. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	9	599
subtype1	3	249
subtype2	3	202
subtype3	3	148

Figure S36. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'Copy Number Ratio CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0165 (Fisher's exact test), Q value = 0.059

Table S40. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	439	5	36	29
subtype1	185	3	8	17
subtype2	153	0	13	6
subtype3	101	2	15	6

Figure S37. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.00164 (Kruskal-Wallis (anova)), Q value = 0.0099

Table S41. Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	573	2.2 (4.7)
subtype1	236	1.8 (4.2)
subtype2	194	1.9 (3.4)
subtype3	143	3.3 (6.5)

Figure S38. Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.722 (Fisher's exact test), Q value = 0.81

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

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	60	292
subtype1	1	7	25	127
subtype2	0	2	23	93
subtype3	0	3	12	72

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

Clustering Approach #4: 'METHLYATION CNMF'

Table S43. Description of clustering approach #4: 'METHLYATION CNMF'

Cluster Labels	1	2	3
Number of samples	94	164	128

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.938 (logrank test), Q value = 0.94

Table S44. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	377	74	0.1 - 140.4 (17.8)
subtype1	93	15	0.1 - 102.4 (15.0)
subtype2	161	35	0.1 - 139.2 (20.6)
subtype3	123	24	0.1 - 140.4 (17.8)

Figure S40. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.00177 (Kruskal-Wallis (anova)), Q value = 0.01

Table S45. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	384	64.4 (13.0)
subtype1	92	60.5 (13.2)
subtype2	164	64.7 (12.3)
subtype3	128	66.8 (13.0)

Figure S41. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.00098 (Fisher's exact test), Q value = 0.0064

Table S46. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	288	96
subtype1	68	25
subtype2	110	53
subtype3	110	18

Figure S42. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'METHLYATION CNMF' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.679 (Fisher's exact test), Q value = 0.78

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

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	54	1	20	112	7	3	11	18	57	35	27	26	2
subtype1	11	0	5	28	1	0	1	7	13	11	6	7	0
subtype2	19	0	9	44	2	2	4	9	29	13	12	13	1
subtype3	24	1	6	40	4	1	6	2	15	11	9	6	1

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

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.568 (Fisher's exact test), Q value = 0.69

Table S48. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	11	57	267	49
subtype1	3	9	67	15
subtype2	4	24	116	19
subtype3	4	24	84	15

Figure S44. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.225 (Fisher's exact test), Q value = 0.37

Table S49. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	210	102	70
subtype1	49	27	18
subtype2	82	50	28
subtype3	79	25	24

Figure S45. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.402 (Fisher's exact test), Q value = 0.55

Table S50. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	262	51
subtype1	67	11
subtype2	106	26
subtype3	89	14

Figure S46. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.0361 (Fisher's exact test), Q value = 0.1

Table S51. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	177	209
subtype1	36	58
subtype2	71	93
subtype3	70	58

Figure S47. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #8: 'GENDER'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.00011 (Fisher's exact test), Q value = 0.001

Table S52. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	250	38	90	6
subtype1	55	13	24	1
subtype2	103	7	51	2
subtype3	92	18	15	3

Figure S48. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.0592 (Fisher's exact test), Q value = 0.14

Table S53. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	8	378
subtype1	4	90
subtype2	4	160
subtype3	0	128

Figure S49. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'METHLYATION CNMF' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.319 (Fisher's exact test), Q value = 0.49

Table S54. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	253	4	6	29
subtype1	64	1	0	5
subtype2	108	1	2	10
subtype3	81	2	4	14

Figure S50. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'METHLYATION CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.562 (Kruskal-Wallis (anova)), Q value = 0.69

Table S55. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	351	2.3 (4.8)
subtype1	85	2.7 (5.2)
subtype2	151	2.2 (5.2)
subtype3	115	2.0 (3.9)

Figure S51. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 0.382 (Fisher's exact test), Q value = 0.54

Table S56. Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	58	282
subtype1	1	4	12	73
subtype2	0	3	23	123
subtype3	0	5	23	86

Figure S52. Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

Clustering Approach #5: 'RPPA CNMF subtypes'

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

Cluster Labels	1	2	3	4
Number of samples	100	179	42	140

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.432 (logrank test), Q value = 0.58

Table S58. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	444	91	0.0 - 140.4 (19.8)
subtype1	94	21	0.0 - 117.1 (18.5)
subtype2	169	39	0.1 - 140.4 (20.0)
subtype3	42	5	0.1 - 62.0 (21.0)
subtype4	139	26	0.0 - 135.7 (19.7)

Figure S53. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.955 (Kruskal-Wallis (anova)), Q value = 0.96

Table S59. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	460	66.8 (12.6)
subtype1	100	67.0 (12.7)
subtype2	178	67.1 (12.5)
subtype3	42	67.4 (10.6)
subtype4	140	66.1 (13.5)

Figure S54. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA CNMF subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.826 (Fisher's exact test), Q value = 0.86

Table S60. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	331	130
subtype1	70	30
subtype2	126	53
subtype3	32	10
subtype4	103	37

Figure S55. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'RPPA CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.0995 (Fisher's exact test), Q value = 0.21

Table S61. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	70	1	27	139	9	2	24	16	62	39	44	19	1
subtype1	10	0	8	26	3	0	3	4	18	10	10	7	0
subtype2	27	0	11	44	4	1	11	7	27	10	26	7	1
subtype3	8	1	3	17	0	0	2	0	3	5	1	1	0
subtype4	25	0	5	52	2	1	8	5	14	14	7	4	0

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.337 (Fisher's exact test), Q value = 0.5

Table S62. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	11	76	324	48
subtype1	0	13	72	14
subtype2	6	31	123	19
subtype3	0	8	32	1
subtype4	5	24	97	14

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0727 (Fisher's exact test), Q value = 0.16

Table S63. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	260	117	81
subtype1	47	33	19
subtype2	95	50	34
subtype3	29	6	6
subtype4	89	28	22

Figure S58. Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0117 (Fisher's exact test), Q value = 0.051

Table S64. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	357	63
subtype1	78	17
subtype2	130	33
subtype3	37	2
subtype4	112	11

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

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.716 (Fisher's exact test), Q value = 0.81

Table S65. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	215	246
subtype1	42	58
subtype2	88	91
subtype3	19	23
subtype4	66	74

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

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.182 (Fisher's exact test), Q value = 0.33

Table S66. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	293	38	117	10
subtype1	61	9	24	4
subtype2	119	7	48	4
subtype3	27	5	9	1
subtype4	86	17	36	1

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

'RPPA CNMF subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.731 (Fisher's exact test), Q value = 0.81

Table S67. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	8	453
subtype1	3	97
subtype2	3	176
subtype3	0	42
subtype4	2	138

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

'RPPA CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0171 (Fisher's exact test), Q value = 0.059

Table S68. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	338	2	28	19
subtype1	74	0	7	7
subtype2	114	1	18	6
subtype3	38	0	1	1
subtype4	112	1	2	5

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

'RPPA CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0256 (Kruskal-Wallis (anova)), Q value = 0.082

Table S69. Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	435	2.1 (4.6)
subtype1	93	2.3 (3.8)
subtype2	169	2.3 (4.4)
subtype3	41	1.5 (3.1)
subtype4	132	2.0 (5.6)

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

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.0461 (Fisher's exact test), Q value = 0.12

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

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	24	246
subtype1	0	6	8	45
subtype2	0	5	8	88
subtype3	0	0	2	12
subtype4	1	1	6	101

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

Clustering Approach #6: 'RPPA cHierClus subtypes'

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

Cluster Labels	1	2	3
Number of samples	54	307	100

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.26 (logrank test), Q value = 0.42

Table S72. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	444	91	0.0 - 140.4 (19.8)
subtype1	53	8	0.1 - 130.7 (17.2)
subtype2	292	68	0.0 - 140.4 (20.0)
subtype3	99	15	0.0 - 135.7 (19.7)

Figure S66. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.65 (Kruskal-Wallis (anova)), Q value = 0.76

Table S73. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	460	66.8 (12.6)
subtype1	54	66.5 (11.9)
subtype2	306	67.1 (12.6)
subtype3	100	65.8 (13.2)

Figure S67. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA cHierClus subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.768 (Fisher's exact test), Q value = 0.83

Table S74. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	331	130
subtype1	39	15
subtype2	223	84
subtype3	69	31

Figure S68. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'RPPA cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.0296 (Fisher's exact test), Q value = 0.091

Table S75. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	70	1	27	139	9	2	24	16	62	39	44	19	1
subtype1	6	0	1	15	1	0	1	3	10	5	4	8	0
subtype2	40	1	22	90	7	1	19	8	41	27	36	8	1
subtype3	24	0	4	34	1	1	4	5	11	7	4	3	0

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

'RPPA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0426 (Fisher's exact test), Q value = 0.12

Table S76. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	11	76	324	48
subtype1	0	7	36	11
subtype2	7	45	223	30
subtype3	4	24	65	7

Figure S70. Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.017 (Fisher's exact test), Q value = 0.059

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

nPatients	N0	N1	N2
ALL	260	117	81
subtype1	23	17	14
subtype2	168	82	55
subtype3	69	18	12

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

'RPPA cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0307 (Fisher's exact test), Q value = 0.092

Table S78. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	357	63
subtype1	37	12
subtype2	239	44
subtype3	81	7

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

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.325 (Fisher's exact test), Q value = 0.49

Table S79. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	215	246
subtype1	26	28
subtype2	149	158
subtype3	40	60

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

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.207 (Fisher's exact test), Q value = 0.35

Table S80. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	293	38	117	10
subtype1	33	6	13	2
subtype2	202	21	73	8
subtype3	58	11	31	0

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

'RPPA cHierClus subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.0578 (Fisher's exact test), Q value = 0.14

Table S81. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	8	453
subtype1	3	51
subtype2	3	304
subtype3	2	98

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

'RPPA cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0147 (Fisher's exact test), Q value = 0.057

Table S82. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	338	2	28	19
subtype1	44	1	1	1
subtype2	212	1	26	14
subtype3	82	0	1	4

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

'RPPA cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.000274 (Kruskal-Wallis (anova)), Q value = 0.0021

Table S83. Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	435	2.1 (4.6)
subtype1	51	3.9 (8.1)
subtype2	290	2.0 (3.9)
subtype3	94	1.3 (3.6)

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

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.76 (Fisher's exact test), Q value = 0.83

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

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	24	246
subtype1	0	2	3	39
subtype2	1	9	15	137
subtype3	0	1	6	70

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

Clustering Approach #7: 'RNAseq CNMF subtypes'

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

Cluster Labels	1	2	3	4
Number of samples	145	142	130	199

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.262 (logrank test), Q value = 0.42

Table S86. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	592	115	0.0 - 140.4 (19.2)
subtype1	140	34	0.1 - 140.4 (19.8)
subtype2	138	23	0.0 - 135.7 (18.7)
subtype3	126	29	0.1 - 131.5 (16.0)
subtype4	188	29	0.0 - 117.1 (21.5)

Figure S79. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00149 (Kruskal-Wallis (anova)), Q value = 0.0093

Table S87. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	614	66.4 (12.7)
subtype1	145	63.5 (12.2)
subtype2	141	66.4 (13.4)
subtype3	129	65.7 (13.1)
subtype4	199	68.8 (11.9)

Figure S80. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.0866 (Fisher's exact test), Q value = 0.19

Table S88. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	449	163
subtype1	100	45
subtype2	110	30
subtype3	102	28
subtype4	137	60

Figure S81. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'RNAseq CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00093 (Fisher's exact test), Q value = 0.0063

Table S89. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	104	1	37	176	11	3	27	21	79	54	60	27	2
subtype1	23	0	7	35	2	2	4	4	24	11	13	13	0
subtype2	29	1	11	42	2	1	8	7	17	6	7	7	2
subtype3	12	0	5	44	2	0	2	5	16	19	15	6	0
subtype4	40	0	14	55	5	0	13	5	22	18	25	1	0

Figure S82. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0514 (Fisher's exact test), Q value = 0.13

Table S90. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	20	107	419	68
subtype1	6	26	95	17
subtype2	4	30	94	13
subtype3	1	13	94	22
subtype4	9	38	136	16

Figure S83. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.155 (Fisher's exact test), Q value = 0.29

Table S91. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	349	149	114
subtype1	80	39	23
subtype2	89	34	19
subtype3	64	31	34
subtype4	116	45	38

Figure S84. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.108 (Fisher's exact test), Q value = 0.22

Table S92. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	456	85
subtype1	94	26
subtype2	101	13
subtype3	91	20
subtype4	170	26

Figure S85. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.654 (Fisher's exact test), Q value = 0.76

Table S93. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	288	328
subtype1	64	81
subtype2	68	74
subtype3	57	73
subtype4	99	100

Figure S86. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'GENDER'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 2e-05 (Fisher's exact test), Q value = 0.00022

Table S94. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	387	61	147	13
subtype1	97	3	45	0
subtype2	86	23	24	5
subtype3	84	18	25	3
subtype4	120	17	53	5

Figure S87. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RNAseq CNMF subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.376 (Fisher's exact test), Q value = 0.53

Table S95. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	9	607
subtype1	2	143
subtype2	4	138
subtype3	2	128
subtype4	1	198

Figure S88. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 2e-05 (Fisher's exact test), Q value = 0.00022

Table S96. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	446	5	35	30
subtype1	87	1	7	10
subtype2	94	2	3	9
subtype3	94	1	3	11
subtype4	171	1	22	0

Figure S89. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RNAseq CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.068 (Kruskal-Wallis (anova)), Q value = 0.15

Table S97. Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	580	2.2 (4.7)
subtype1	131	1.9 (3.9)
subtype2	132	1.8 (4.1)
subtype3	120	3.2 (6.5)
subtype4	197	2.1 (4.3)

Figure S90. Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.782 (Fisher's exact test), Q value = 0.84

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

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	60	294
subtype1	0	4	18	96
subtype2	0	4	22	77
subtype3	1	4	16	101
subtype4	0	0	4	20

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

Clustering Approach #8: 'RNAseq cHierClus subtypes'

Table S99. Description of clustering approach #8: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3	4	5	6	7
Number of samples	79	84	63	57	90	140	103

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.533 (logrank test), Q value = 0.67

Table S100. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	592	115	0.0 - 140.4 (19.2)
subtype1	77	19	0.1 - 140.4 (17.5)
subtype2	81	14	0.8 - 124.3 (17.8)
subtype3	61	14	0.1 - 135.7 (19.4)
subtype4	55	7	0.5 - 91.8 (20.0)
subtype5	90	19	0.3 - 129.3 (15.0)
subtype6	130	20	0.0 - 117.1 (21.5)
subtype7	98	22	0.0 - 64.0 (24.0)

Figure S92. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.000798 (Kruskal-Wallis (anova)), Q value = 0.0057

Table S101. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	614	66.4 (12.7)
subtype1	79	64.3 (13.0)
subtype2	84	64.7 (13.4)
subtype3	61	65.1 (15.2)
subtype4	57	64.7 (12.8)
subtype5	90	63.9 (11.6)
subtype6	140	68.0 (10.9)
subtype7	103	70.8 (12.2)

Figure S93. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S102. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	449	163
subtype1	57	22
subtype2	71	13
subtype3	59	4
subtype4	32	23
subtype5	59	31
subtype6	88	51
subtype7	83	19

Figure S94. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'RNAseq cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S103. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	104	1	37	176	11	3	27	21	79	54	60	27	2
subtype1	14	0	1	24	1	1	0	4	12	9	5	5	0
subtype2	15	1	6	23	1	0	3	4	14	5	4	5	1
subtype3	13	0	4	28	1	0	2	1	4	5	1	1	1
subtype4	8	0	3	8	1	1	5	5	9	3	6	6	0
subtype5	7	0	3	26	1	1	1	2	15	12	9	9	0
subtype6	28	0	7	37	1	0	8	4	16	14	25	0	0
subtype7	19	0	13	30	5	0	8	1	9	6	10	1	0

Figure S95. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.613 (Fisher's exact test), Q value = 0.74

Table S104. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	20	107	419	68
subtype1	2	14	54	9
subtype2	5	16	50	12
subtype3	2	11	44	6
subtype4	1	9	40	7
subtype5	0	10	66	13
subtype6	6	30	94	10
subtype7	4	17	71	11

Figure S96. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0138 (Fisher's exact test), Q value = 0.055

Table S105. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	349	149	114
subtype1	43	18	17
subtype2	50	22	11
subtype3	46	9	8
subtype4	26	22	9
subtype5	39	27	22
subtype6	77	34	29
subtype7	68	17	18

Figure S97. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0511 (Fisher's exact test), Q value = 0.13

Table S106. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	456	85
subtype1	55	10
subtype2	56	7
subtype3	46	3
subtype4	34	11
subtype5	61	18
subtype6	113	25
subtype7	91	11

Figure S98. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.346 (Fisher's exact test), Q value = 0.5

Table S107. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	288	328
subtype1	37	42
subtype2	41	43
subtype3	28	35
subtype4	18	39
subtype5	43	47
subtype6	68	72
subtype7	53	50

Figure S99. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S108. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	387	61	147	13
subtype1	54	3	22	0
subtype2	57	14	11	2
subtype3	47	12	3	1
subtype4	30	2	22	1
subtype5	53	6	30	1
subtype6	86	2	50	0
subtype7	60	22	9	8

Figure S100. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.0407 (Fisher's exact test), Q value = 0.11

Table S109. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	9	607
subtype1	0	79
subtype2	1	83
subtype3	1	62
subtype4	3	54
subtype5	3	87
subtype6	0	140
subtype7	1	102

Figure S101. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S110. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	446	5	35	30
subtype1	44	0	2	8
subtype2	56	0	1	4
subtype3	41	1	0	9
subtype4	34	2	2	2
subtype5	66	0	0	7
subtype6	115	2	22	0
subtype7	90	0	8	0

Figure S102. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RNAseq cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0134 (Kruskal-Wallis (anova)), Q value = 0.055

Table S111. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	580	2.2 (4.7)
subtype1	71	2.2 (3.2)
subtype2	77	1.8 (4.5)
subtype3	56	1.3 (3.1)
subtype4	53	2.6 (5.1)
subtype5	82	3.1 (6.5)
subtype6	138	2.2 (4.4)
subtype7	103	2.2 (5.1)

Figure S103. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.1 (Fisher's exact test), Q value = 0.21

Table S112. Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	60	294
subtype1	0	3	12	59
subtype2	0	0	21	49
subtype3	0	4	7	49
subtype4	0	1	4	43
subtype5	1	4	11	71
subtype6	0	0	4	13
subtype7	0	0	1	10

Figure S104. Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #9: 'MIRSEQ CNMF'

Table S113. Description of clustering approach #9: 'MIRSEQ CNMF'

Cluster Labels	1	2	3
Number of samples	212	96	241

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0468 (logrank test), Q value = 0.12

Table S114. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	525	107	0.0 - 140.4 (20.0)
subtype1	205	40	0.1 - 140.4 (18.3)
subtype2	95	23	0.1 - 100.0 (15.0)
subtype3	225	44	0.0 - 117.1 (22.4)

Figure S105. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.00875 (Kruskal-Wallis (anova)), Q value = 0.041

Table S115. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	548	66.8 (12.7)
subtype1	211	65.9 (12.5)
subtype2	96	64.0 (13.8)
subtype3	241	68.8 (12.1)

Figure S106. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.0403 (Fisher's exact test), Q value = 0.11

Table S116. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	405	140
subtype1	150	61
subtype2	80	15
subtype3	175	64

Figure S107. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'MIRSEQ CNMF' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00021 (Fisher's exact test), Q value = 0.0017

Table S117. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	94	1	36	154	10	2	27	16	69	45	57	23	1
subtype1	32	1	8	67	4	2	8	9	33	13	16	11	1
subtype2	9	0	5	26	0	0	3	3	15	14	7	8	0
subtype3	53	0	23	61	6	0	16	4	21	18	34	4	0

Figure S108. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.104 (Fisher's exact test), Q value = 0.21

Table S118. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	20	95	376	56
subtype1	8	33	151	18
subtype2	2	10	69	15
subtype3	10	52	156	23

Figure S109. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.0169 (Fisher's exact test), Q value = 0.059

Table S119. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	312	133	100
subtype1	120	58	31
subtype2	43	29	23
subtype3	149	46	46

Figure S110. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.816 (Fisher's exact test), Q value = 0.86

Table S120. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	412	78
subtype1	149	26
subtype2	65	14
subtype3	198	38

Figure S111. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.541 (Fisher's exact test), Q value = 0.67

Table S121. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	257	292
subtype1	102	110
subtype2	40	56
subtype3	115	126

Figure S112. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #8: 'GENDER'

'MIRSEQ CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.00031 (Fisher's exact test), Q value = 0.0023

Table S122. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	351	53	127	10
subtype1	140	10	60	1
subtype2	61	19	13	2
subtype3	150	24	54	7

Figure S113. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CNMF' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.0948 (Fisher's exact test), Q value = 0.2

Table S123. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	9	540
subtype1	3	209
subtype2	4	92
subtype3	2	239

Figure S114. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'MIRSEQ CNMF' versus 'COMPLETENESS_OF_RESECTION'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S124. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	384	5	36	26
subtype1	121	2	5	17
subtype2	66	0	2	7
subtype3	197	3	29	2

Figure S115. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRSEQ CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0488 (Kruskal-Wallis (anova)), Q value = 0.12

Table S125. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	514	2.2 (4.7)
subtype1	189	1.8 (3.5)
subtype2	89	3.3 (7.1)
subtype3	236	2.1 (4.4)

Figure S116. Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.661 (Fisher's exact test), Q value = 0.76

Table S126. Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	26	280
subtype1	0	7	14	171
subtype2	1	4	7	73
subtype3	0	1	5	36

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

Clustering Approach #10: 'MIRSEQ CHIERARCHICAL'

Table S127. Description of clustering approach #10: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	229	144	176

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.843 (logrank test), Q value = 0.87

Table S128. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	525	107	0.0 - 140.4 (20.0)
subtype1	224	49	0.1 - 140.4 (17.4)
subtype2	137	27	0.0 - 139.2 (19.1)
subtype3	164	31	0.0 - 83.8 (25.0)

Figure S118. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.00428 (Kruskal-Wallis (anova)), Q value = 0.022

Table S129. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	548	66.8 (12.7)
subtype1	228	65.8 (12.8)
subtype2	144	65.1 (13.0)
subtype3	176	69.6 (11.9)

Figure S119. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 0.263 (Fisher's exact test), Q value = 0.42

Table S130. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	405	140
subtype1	176	51
subtype2	100	43
subtype3	129	46

Figure S120. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00974 (Fisher's exact test), Q value = 0.045

Table S131. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA	STAGE IVB
ALL	94	1	36	154	10	2	27	16	69	45	57	23	1
subtype1	36	0	10	78	4	1	8	8	28	18	13	14	1
subtype2	22	1	6	34	2	1	7	6	21	14	23	4	0
subtype3	36	0	20	42	4	0	12	2	20	13	21	5	0

Figure S121. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.92 (Fisher's exact test), Q value = 0.93

Table S132. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	20	95	376	56
subtype1	8	37	158	25
subtype2	6	22	101	14
subtype3	6	36	117	17

Figure S122. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 0.144 (Fisher's exact test), Q value = 0.27

Table S133. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	312	133	100
subtype1	137	53	36
subtype2	69	42	32
subtype3	106	38	32

Figure S123. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.192 (Fisher's exact test), Q value = 0.34

Table S134. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	412	78
subtype1	163	25
subtype2	103	27
subtype3	146	26

Figure S124. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.519 (Fisher's exact test), Q value = 0.66

Table S135. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	257	292
subtype1	101	128
subtype2	72	72
subtype3	84	92

Figure S125. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL_TYPE'

P value = 0.339 (Fisher's exact test), Q value = 0.5

Table S136. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	351	53	127	10
subtype1	151	25	49	2
subtype2	90	9	40	3
subtype3	110	19	38	5

Figure S126. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.4 (Fisher's exact test), Q value = 0.55

Table S137. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	9	540
subtype1	6	223
subtype2	1	143
subtype3	2	174

Figure S127. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'MIRSEQ CHIERARCHICAL' versus 'COMPLETENESS_OF_RESECTION'

P value = 8e-05 (Fisher's exact test), Q value = 0.00078

Table S138. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	384	5	36	26
subtype1	142	2	4	17
subtype2	95	1	15	7
subtype3	147	2	17	2

Figure S128. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.127 (Kruskal-Wallis (anova)), Q value = 0.25

Table S139. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	514	2.2 (4.7)
subtype1	212	1.9 (3.9)
subtype2	129	2.5 (5.4)
subtype3	173	2.3 (5.2)

Figure S129. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.79 (Fisher's exact test), Q value = 0.84

Table S140. Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	12	26	280
subtype1	1	8	15	181
subtype2	0	4	7	68
subtype3	0	0	4	31

Figure S130. Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

Clustering Approach #11: 'MIRseq Mature CNMF subtypes'

Table S141. Description of clustering approach #11: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3	4	5	6
Number of samples	24	20	29	28	18	17

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

P value = 0.72 (logrank test), Q value = 0.81

Table S142. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	133	37	0.1 - 140.4 (17.0)
subtype1	23	6	0.2 - 130.7 (24.1)
subtype2	20	6	0.1 - 67.3 (15.6)
subtype3	28	6	0.4 - 129.3 (15.4)
subtype4	27	10	0.5 - 140.4 (13.1)
subtype5	18	5	2.0 - 139.2 (15.5)
subtype6	17	4	0.1 - 131.5 (21.4)

Figure S131. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.201 (Kruskal-Wallis (anova)), Q value = 0.34

Table S143. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	135	64.9 (12.6)
subtype1	24	68.0 (12.5)
subtype2	20	62.6 (14.2)
subtype3	29	61.9 (12.3)
subtype4	28	66.4 (10.9)
subtype5	18	62.4 (12.8)
subtype6	16	68.6 (13.2)

Figure S132. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S144. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	85	50
subtype1	13	11
subtype2	20	0
subtype3	6	22
subtype4	16	12
subtype5	16	2
subtype6	14	3

Figure S133. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'MIRseq Mature CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.196 (Fisher's exact test), Q value = 0.34

Table S145. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	15	7	39	1	1	2	4	21	14	13	9
subtype1	6	1	7	0	0	1	0	4	1	2	1
subtype2	2	1	4	0	0	0	0	4	4	4	1
subtype3	0	1	8	0	0	0	1	5	4	4	4
subtype4	1	0	10	0	1	1	3	4	2	0	1
subtype5	2	2	4	1	0	0	0	1	1	3	2
subtype6	4	2	6	0	0	0	0	3	2	0	0

Figure S134. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.144 (Fisher's exact test), Q value = 0.27

Table S146. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	4	16	100	15
subtype1	2	5	15	2
subtype2	0	2	16	2
subtype3	0	0	26	2
subtype4	1	3	21	3
subtype5	0	3	10	5
subtype6	1	3	12	1

Figure S135. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.217 (Fisher's exact test), Q value = 0.36

Table S147. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	65	39	30
subtype1	13	6	5
subtype2	7	5	8
subtype3	9	10	8
subtype4	13	10	5
subtype5	10	6	2
subtype6	13	2	2

Figure S136. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.0257 (Fisher's exact test), Q value = 0.082

Table S148. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	87	21
subtype1	17	2
subtype2	11	5
subtype3	17	7
subtype4	18	2
subtype5	8	5
subtype6	16	0

Figure S137. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.856 (Fisher's exact test), Q value = 0.88

Table S149. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	58	78
subtype1	8	16
subtype2	9	11
subtype3	15	14
subtype4	12	16
subtype5	7	11
subtype6	7	10

Figure S138. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S150. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	75	10	49	1
subtype1	13	0	11	0
subtype2	16	4	0	0
subtype3	6	0	22	0
subtype4	14	2	11	1
subtype5	12	4	2	0
subtype6	14	0	3	0

Figure S139. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.879 (Fisher's exact test), Q value = 0.9

Table S151. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	4	132
subtype1	1	23
subtype2	0	20
subtype3	2	27
subtype4	1	27
subtype5	0	18
subtype6	0	17

Figure S140. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'MIRseq Mature CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.535 (Fisher's exact test), Q value = 0.67

Table S152. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	73	1	2	7
subtype1	12	0	0	2
subtype2	14	0	0	0
subtype3	18	0	2	2
subtype4	11	1	0	0
subtype5	9	0	0	1
subtype6	9	0	0	2

Figure S141. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0318 (Kruskal-Wallis (anova)), Q value = 0.093

Table S153. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	125	2.9 (6.2)
subtype1	23	2.6 (4.4)
subtype2	19	6.9 (11.6)
subtype3	25	3.1 (4.7)
subtype4	26	2.0 (5.4)
subtype5	17	1.4 (3.0)
subtype6	15	1.1 (2.3)

Figure S142. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.219 (Fisher's exact test), Q value = 0.36

Table S154. Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	3	8	117
subtype1	0	0	2	20
subtype2	1	0	2	16
subtype3	0	1	0	25
subtype4	0	0	1	27
subtype5	0	0	2	15
subtype6	0	2	1	14

Figure S143. Get High-res Image Clustering Approach #11: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

Clustering Approach #12: 'MIRseq Mature cHierClus subtypes'

Table S155. Description of clustering approach #12: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	55	15	34	32

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

P value = 0.565 (logrank test), Q value = 0.69

Table S156. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	133	37	0.1 - 140.4 (17.0)
subtype1	55	18	0.1 - 100.0 (16.2)
subtype2	15	3	0.7 - 129.3 (14.5)
subtype3	33	9	0.2 - 131.5 (24.4)
subtype4	30	7	0.4 - 140.4 (15.0)

Figure S144. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.516 (Kruskal-Wallis (anova)), Q value = 0.66

Table S157. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	135	64.9 (12.6)
subtype1	55	64.6 (13.8)
subtype2	15	61.7 (12.3)
subtype3	33	67.3 (12.6)
subtype4	32	64.3 (10.5)

Figure S145. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PRIMARY_SITE_OF_DISEASE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S158. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

nPatients	COLON	RECTUM
ALL	85	50
subtype1	50	5
subtype2	5	9
subtype3	18	16
subtype4	12	20

Figure S146. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PRIMARY_SITE_OF_DISEASE'

'MIRseq Mature cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.767 (Fisher's exact test), Q value = 0.83

Table S159. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	15	7	39	1	1	2	4	21	14	13	9
subtype1	9	3	15	1	0	0	0	7	7	6	2
subtype2	0	1	5	0	0	0	1	4	1	1	2
subtype3	4	2	9	0	1	1	2	7	3	2	1
subtype4	2	1	10	0	0	1	1	3	3	4	4

Figure S147. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'NEOPLASM_DISEASESTAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.44 (Fisher's exact test), Q value = 0.58

Table S160. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	4	16	100	15
subtype1	1	9	36	9
subtype2	0	0	14	1
subtype3	2	5	25	2
subtype4	1	2	25	3

Figure S148. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.751 (Fisher's exact test), Q value = 0.83

Table S161. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2
ALL	65	39	30
subtype1	31	12	12
subtype2	6	5	4
subtype3	14	12	8
subtype4	14	10	6

Figure S149. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.471 (Fisher's exact test), Q value = 0.61

Table S162. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	87	21
subtype1	34	8
subtype2	10	2
subtype3	23	3
subtype4	20	8

Figure S150. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY_M_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.323 (Fisher's exact test), Q value = 0.49

Table S163. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

nPatients	FEMALE	MALE
ALL	58	78
subtype1	23	32
subtype2	9	6
subtype3	11	23
subtype4	15	17

Figure S151. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.00014

Table S164. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	COLON ADENOCARCINOMA	COLON MUCINOUS ADENOCARCINOMA	RECTAL ADENOCARCINOMA	RECTAL MUCINOUS ADENOCARCINOMA
ALL	75	10	49	1
subtype1	42	8	5	0
subtype2	5	0	9	0
subtype3	18	0	15	1
subtype4	10	2	20	0

Figure S152. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATIONS_RADIATION_REGIMENINDICATION'

P value = 0.144 (Fisher's exact test), Q value = 0.27

Table S165. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

nPatients	NO	YES
ALL	4	132
subtype1	0	55
subtype2	1	14
subtype3	2	32
subtype4	1	31

Figure S153. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RADIATIONS_RADIATION_REGIMENINDICATION'

'MIRseq Mature cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.356 (Fisher's exact test), Q value = 0.51

Table S166. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	73	1	2	7
subtype1	33	0	0	3
subtype2	11	0	0	1
subtype3	13	1	0	2
subtype4	16	0	2	1

Figure S154. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.348 (Kruskal-Wallis (anova)), Q value = 0.5

Table S167. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	125	2.9 (6.2)
subtype1	51	3.0 (7.7)
subtype2	14	3.6 (5.8)
subtype3	32	3.3 (5.8)
subtype4	28	1.8 (2.9)

Figure S155. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.796 (Fisher's exact test), Q value = 0.84

Table S168. Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	3	8	117
subtype1	1	1	5	46
subtype2	0	1	0	13
subtype3	0	1	2	30
subtype4	0	0	1	28

Figure S156. Get High-res Image Clustering Approach #12: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/COADREAD-TP/15117685/COADREAD-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/COADREAD-TP/15076882/COADREAD-TP.merged_data.txt
Number of patients = 621
Number of clustering approaches = 12
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)

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