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 15 clinical features across 185 patients, 18 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 'GENDER'.
4 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time to Death', 'PATHOLOGY_T_STAGE', and 'GENDER'.
CNMF clustering analysis on RPPA data identified 5 subtypes that correlate to 'GENDER'.
Consensus hierarchical clustering analysis on RPPA data identified 5 subtypes that correlate to 'GENDER'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'PATHOLOGIC_STAGE' and 'PATHOLOGY_T_STAGE'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time to Death', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', and 'HISTOLOGICAL_TYPE'.
4 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'HISTOLOGICAL_TYPE'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'PATHOLOGIC_STAGE' and 'PATHOLOGY_T_STAGE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes do not correlate to any clinical features.
3 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'PATHOLOGIC_STAGE' and 'PATHOLOGY_T_STAGE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 10 different clustering approaches and 15 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 18 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
Time to Death	logrank test	0.255 (0.552)	0.00998 (0.0989)	0.556 (0.794)	0.224 (0.552)	0.178 (0.552)	0.0106 (0.0989)	0.19 (0.552)	0.11 (0.478)	0.073 (0.438)	0.0458 (0.316)
YEARS TO BIRTH	Kruskal-Wallis (anova)	0.996 (1.00)	0.222 (0.552)	0.278 (0.552)	0.615 (0.828)	0.357 (0.6)	0.174 (0.552)	0.566 (0.798)	0.577 (0.798)	0.519 (0.763)	0.27 (0.552)
PATHOLOGIC STAGE	Fisher's exact test	0.0383 (0.302)	0.14 (0.539)	0.311 (0.575)	0.226 (0.552)	0.00178 (0.0445)	5e-05 (0.0065)	0.0557 (0.363)	0.0067 (0.0773)	0.0791 (0.441)	0.00367 (0.0612)
PATHOLOGY T STAGE	Fisher's exact test	0.0794 (0.441)	0.00076 (0.0228)	0.259 (0.552)	0.306 (0.575)	0.00013 (0.0065)	0.0001 (0.0065)	0.0463 (0.316)	0.00075 (0.0228)	0.146 (0.539)	0.0029 (0.0591)
PATHOLOGY N STAGE	Fisher's exact test	0.917 (0.969)	0.215 (0.552)	0.262 (0.552)	0.28 (0.552)	0.206 (0.552)	0.00315 (0.0591)	0.353 (0.6)	0.111 (0.478)	0.883 (0.946)	0.205 (0.552)
PATHOLOGY M STAGE	Fisher's exact test	0.791 (0.912)	0.247 (0.552)	0.36 (0.6)	0.296 (0.569)	0.27 (0.552)	0.103 (0.478)	0.188 (0.552)	0.339 (0.597)	0.469 (0.704)	0.669 (0.841)
GENDER	Fisher's exact test	0.0246 (0.205)	0.00653 (0.0773)	0.00435 (0.0652)	0.023 (0.203)	0.674 (0.841)	0.979 (1.00)	0.695 (0.841)	0.315 (0.577)	0.301 (0.572)	0.516 (0.763)
RADIATION THERAPY	Fisher's exact test	0.388 (0.633)	0.524 (0.764)	0.102 (0.478)	0.676 (0.841)	0.329 (0.588)	0.67 (0.841)	0.183 (0.552)	0.64 (0.841)	0.232 (0.552)	0.639 (0.841)
HISTOLOGICAL TYPE	Fisher's exact test	0.0668 (0.418)	0.243 (0.552)	0.17 (0.552)	0.24 (0.552)	0.108 (0.478)	0.0087 (0.0932)	0.00652 (0.0773)	0.0423 (0.316)	0.328 (0.588)	0.218 (0.552)
NUMBER PACK YEARS SMOKED	Kruskal-Wallis (anova)	0.351 (0.6)	0.784 (0.911)	0.291 (0.567)	0.963 (1.00)	0.218 (0.552)	0.205 (0.552)	0.543 (0.784)	0.23 (0.552)	0.153 (0.539)	0.253 (0.552)
YEAR OF TOBACCO SMOKING ONSET	Kruskal-Wallis (anova)	0.269 (0.552)	0.82 (0.932)	0.147 (0.539)	0.65 (0.841)	0.767 (0.899)	0.908 (0.966)	0.218 (0.552)	0.827 (0.933)	0.986 (1.00)	0.58 (0.798)
RESIDUAL TUMOR	Fisher's exact test	0.111 (0.478)	0.572 (0.798)	0.949 (0.996)	0.434 (0.681)	0.243 (0.552)	0.146 (0.539)	0.619 (0.828)	0.755 (0.892)	0.679 (0.841)	0.46 (0.703)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.154 (0.539)	0.466 (0.704)	0.436 (0.681)	0.152 (0.539)	0.69 (0.841)	0.0856 (0.459)	0.688 (0.841)	0.451 (0.697)	0.729 (0.868)	0.677 (0.841)
RACE	Fisher's exact test	0.342 (0.597)	0.971 (1.00)	0.805 (0.922)	0.844 (0.936)	0.13 (0.539)	0.0989 (0.478)	0.372 (0.613)	0.418 (0.667)	0.276 (0.552)	0.231 (0.552)
ETHNICITY	Fisher's exact test	1 (1.00)	0.868 (0.938)	0.855 (0.936)	1 (1.00)	0.853 (0.936)	0.399 (0.643)	0.869 (0.938)	0.608 (0.828)	0.842 (0.936)	0.723 (0.868)

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	85	75	24

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

P value = 0.255 (logrank test), Q value = 0.55

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	182	96	0.0 - 90.1 (15.2)
subtype1	83	41	0.3 - 75.1 (15.3)
subtype2	75	43	0.0 - 66.9 (12.9)
subtype3	24	12	0.2 - 90.1 (16.6)

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

'Copy Number Ratio CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.996 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	184	64.9 (11.1)
subtype1	85	65.1 (10.7)
subtype2	75	64.8 (11.6)
subtype3	24	64.9 (11.0)

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0383 (Fisher's exact test), Q value = 0.3

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

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	30	121	5	5
subtype1	0	3	7	12	57	3	2
subtype2	0	2	2	17	49	2	3
subtype3	1	0	6	1	15	0	0

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0794 (Fisher's exact test), Q value = 0.44

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

nPatients	T1	T2	T3	T4
ALL	7	24	147	4
subtype1	4	12	65	3
subtype2	2	5	67	1
subtype3	1	7	15	0

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.917 (Fisher's exact test), Q value = 0.97

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

nPatients	0	1
ALL	50	129
subtype1	23	61
subtype2	22	52
subtype3	5	16

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

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

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

nPatients	0	1
ALL	85	5
subtype1	42	2
subtype2	35	3
subtype3	8	0

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

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.0246 (Fisher's exact test), Q value = 0.21

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

nPatients	FEMALE	MALE
ALL	83	101
subtype1	44	41
subtype2	34	41
subtype3	5	19

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

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.388 (Fisher's exact test), Q value = 0.63

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

nPatients	NO	YES
ALL	122	42
subtype1	55	20
subtype2	48	19
subtype3	19	3

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

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.0668 (Fisher's exact test), Q value = 0.42

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

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	153	25	4	1
subtype1	71	10	3	1
subtype2	66	8	0	0
subtype3	16	7	1	0

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.351 (Kruskal-Wallis (anova)), Q value = 0.6

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

	nPatients	Mean (Std.Dev)
ALL	56	26.9 (18.1)
subtype1	27	25.5 (18.9)
subtype2	24	29.8 (17.8)
subtype3	5	20.2 (15.8)

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

'Copy Number Ratio CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.269 (Kruskal-Wallis (anova)), Q value = 0.55

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

	nPatients	Mean (Std.Dev)
ALL	46	1970.4 (13.1)
subtype1	23	1970.0 (14.3)
subtype2	18	1973.1 (12.2)
subtype3	5	1962.6 (7.4)

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

'Copy Number Ratio CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.111 (Fisher's exact test), Q value = 0.48

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

nPatients	R0	R1	R2	RX
ALL	111	53	5	4
subtype1	58	18	2	1
subtype2	40	26	3	1
subtype3	13	9	0	2

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.154 (Kruskal-Wallis (anova)), Q value = 0.54

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

	nPatients	Mean (Std.Dev)
ALL	180	3.0 (3.5)
subtype1	84	3.0 (3.5)
subtype2	74	2.5 (2.8)
subtype3	22	4.7 (4.8)

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

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.342 (Fisher's exact test), Q value = 0.6

Table S15. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	7	161
subtype1	4	5	74
subtype2	7	2	64
subtype3	0	0	23

Figure S14. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #14: 'RACE'

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

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

Table S16. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	136
subtype1	3	68
subtype2	2	48
subtype3	0	20

Figure S15. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #2: 'METHLYATION CNMF'

Table S17. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3	4
Number of samples	45	48	50	41

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.00998 (logrank test), Q value = 0.099

Table S18. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	182	95	0.0 - 90.1 (15.2)
subtype1	45	28	0.0 - 71.7 (15.6)
subtype2	48	18	0.3 - 90.1 (16.0)
subtype3	49	32	0.1 - 66.9 (15.1)
subtype4	40	17	0.3 - 49.4 (13.9)

Figure S16. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.222 (Kruskal-Wallis (anova)), Q value = 0.55

Table S19. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	184	64.8 (11.0)
subtype1	45	67.6 (10.6)
subtype2	48	64.2 (10.2)
subtype3	50	64.0 (11.6)
subtype4	41	63.1 (11.4)

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

'METHLYATION CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.14 (Fisher's exact test), Q value = 0.54

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	30	121	5	5
subtype1	0	2	4	7	29	1	1
subtype2	1	3	8	6	27	1	1
subtype3	0	0	1	12	33	1	3
subtype4	0	0	2	5	32	2	0

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.00076 (Fisher's exact test), Q value = 0.023

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	147	4
subtype1	2	4	36	2
subtype2	5	13	29	0
subtype3	0	3	47	0
subtype4	0	4	35	2

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.215 (Fisher's exact test), Q value = 0.55

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	50	129
subtype1	15	30
subtype2	16	29
subtype3	12	37
subtype4	7	33

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.247 (Fisher's exact test), Q value = 0.55

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	84	5
subtype1	21	1
subtype2	20	1
subtype3	19	3
subtype4	24	0

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.00653 (Fisher's exact test), Q value = 0.077

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	82	102
subtype1	11	34
subtype2	23	25
subtype3	23	27
subtype4	25	16

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.524 (Fisher's exact test), Q value = 0.76

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	122	42
subtype1	31	9
subtype2	31	16
subtype3	34	10
subtype4	26	7

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.243 (Fisher's exact test), Q value = 0.55

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	153	25	4	1
subtype1	40	5	0	0
subtype2	35	11	2	0
subtype3	42	6	1	0
subtype4	36	3	1	1

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.784 (Kruskal-Wallis (anova)), Q value = 0.91

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	57	26.8 (17.9)
subtype1	17	26.3 (12.9)
subtype2	10	24.6 (24.6)
subtype3	14	26.0 (17.1)
subtype4	16	29.6 (19.9)

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'METHLYATION CNMF' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.82 (Kruskal-Wallis (anova)), Q value = 0.93

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	16	1969.8 (12.2)
subtype2	7	1975.4 (16.9)
subtype3	10	1970.5 (13.3)
subtype4	14	1969.6 (13.1)

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'METHLYATION CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.572 (Fisher's exact test), Q value = 0.8

Table S29. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	111	53	5	4
subtype1	24	16	1	2
subtype2	32	12	0	0
subtype3	29	13	2	2
subtype4	26	12	2	0

Figure S27. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'METHLYATION CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.466 (Kruskal-Wallis (anova)), Q value = 0.7

Table S30. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	180	3.0 (3.5)
subtype1	44	3.2 (3.6)
subtype2	46	2.7 (3.8)
subtype3	50	2.9 (3.4)
subtype4	40	3.2 (3.1)

Figure S28. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 0.971 (Fisher's exact test), Q value = 1

Table S31. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	7	162
subtype1	3	2	39
subtype2	2	2	43
subtype3	4	1	45
subtype4	2	2	35

Figure S29. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.868 (Fisher's exact test), Q value = 0.94

Table S32. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	137
subtype1	1	31
subtype2	1	37
subtype3	1	38
subtype4	2	31

Figure S30. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #3: 'RPPA CNMF subtypes'

Table S33. Description of clustering approach #3: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	40	25	20	15	23

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.556 (logrank test), Q value = 0.79

Table S34. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	123	68	0.0 - 71.7 (14.9)
subtype1	40	22	0.9 - 71.7 (14.3)
subtype2	25	16	2.5 - 66.9 (15.3)
subtype3	20	10	0.1 - 68.5 (14.3)
subtype4	15	6	0.0 - 45.5 (17.8)
subtype5	23	14	3.1 - 43.8 (12.0)

Figure S31. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.278 (Kruskal-Wallis (anova)), Q value = 0.55

Table S35. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	123	65.3 (11.5)
subtype1	40	64.9 (11.0)
subtype2	25	67.0 (12.0)
subtype3	20	60.5 (10.8)
subtype4	15	68.1 (9.4)
subtype5	23	66.8 (13.1)

Figure S32. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.311 (Fisher's exact test), Q value = 0.58

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

nPatients	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	5	8	19	83	5	2
subtype1	1	5	7	27	0	0
subtype2	1	0	4	16	3	1
subtype3	1	2	4	12	0	0
subtype4	1	0	0	12	1	1
subtype5	1	1	4	16	1	0

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.259 (Fisher's exact test), Q value = 0.55

Table S37. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	6	11	101	4
subtype1	1	7	32	0
subtype2	1	0	22	2
subtype3	2	2	15	0
subtype4	1	1	12	1
subtype5	1	1	20	1

Figure S34. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.262 (Fisher's exact test), Q value = 0.55

Table S38. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	32	90
subtype1	13	27
subtype2	6	19
subtype3	7	12
subtype4	1	14
subtype5	5	18

Figure S35. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.36 (Fisher's exact test), Q value = 0.6

Table S39. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	60	2
subtype1	21	0
subtype2	11	1
subtype3	8	0
subtype4	8	1
subtype5	12	0

Figure S36. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.00435 (Fisher's exact test), Q value = 0.065

Table S40. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	57	66
subtype1	14	26
subtype2	20	5
subtype3	7	13
subtype4	6	9
subtype5	10	13

Figure S37. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.102 (Fisher's exact test), Q value = 0.48

Table S41. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	82	28
subtype1	24	10
subtype2	16	8
subtype3	11	6
subtype4	11	3
subtype5	20	1

Figure S38. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.17 (Fisher's exact test), Q value = 0.55

Table S42. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA
ALL	108	12	2
subtype1	36	3	1
subtype2	20	3	1
subtype3	18	2	0
subtype4	11	4	0
subtype5	23	0	0

Figure S39. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RPPA CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.291 (Kruskal-Wallis (anova)), Q value = 0.57

Table S43. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	39	26.0 (16.7)
subtype1	15	31.7 (12.0)
subtype2	8	25.1 (21.3)
subtype3	7	20.1 (21.6)
subtype4	6	17.8 (15.3)
subtype5	3	30.0 (10.0)

Figure S40. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.147 (Kruskal-Wallis (anova)), Q value = 0.54

Table S44. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	32	1970.9 (13.4)
subtype1	11	1968.0 (11.5)
subtype2	9	1966.3 (12.0)
subtype3	6	1978.2 (16.0)
subtype4	4	1980.5 (6.6)
subtype5	2	1966.0 (25.5)

Figure S41. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RPPA CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.949 (Fisher's exact test), Q value = 1

Table S45. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	72	39	2	2
subtype1	22	14	1	1
subtype2	12	9	0	0
subtype3	15	5	0	0
subtype4	10	4	0	0
subtype5	13	7	1	1

Figure S42. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RPPA CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.436 (Kruskal-Wallis (anova)), Q value = 0.68

Table S46. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	122	2.9 (3.2)
subtype1	40	2.2 (2.5)
subtype2	25	3.5 (3.9)
subtype3	19	2.5 (2.6)
subtype4	15	3.8 (4.2)
subtype5	23	3.3 (3.4)

Figure S43. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.805 (Fisher's exact test), Q value = 0.92

Table S47. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	4	113
subtype1	1	3	35
subtype2	1	0	21
subtype3	0	0	20
subtype4	0	0	15
subtype5	0	1	22

Figure S44. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.855 (Fisher's exact test), Q value = 0.94

Table S48. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	95
subtype1	2	29
subtype2	0	17
subtype3	0	17
subtype4	0	12
subtype5	1	20

Figure S45. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S49. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	48	25	16	16	18

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.224 (logrank test), Q value = 0.55

Table S50. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	123	68	0.0 - 71.7 (14.9)
subtype1	48	29	2.0 - 71.7 (13.4)
subtype2	25	15	2.5 - 66.9 (15.3)
subtype3	16	7	0.3 - 59.0 (14.1)
subtype4	16	5	0.0 - 68.5 (18.7)
subtype5	18	12	0.1 - 43.8 (12.3)

Figure S46. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.615 (Kruskal-Wallis (anova)), Q value = 0.83

Table S51. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	123	65.3 (11.5)
subtype1	48	65.3 (12.0)
subtype2	25	67.8 (11.6)
subtype3	16	63.0 (10.1)
subtype4	16	66.6 (7.7)
subtype5	18	63.1 (14.1)

Figure S47. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.226 (Fisher's exact test), Q value = 0.55

Table S52. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	5	8	19	83	5	2
subtype1	3	4	8	33	0	0
subtype2	1	0	4	15	4	1
subtype3	0	2	4	9	0	0
subtype4	1	1	1	11	1	1
subtype5	0	1	2	15	0	0

Figure S48. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.306 (Fisher's exact test), Q value = 0.57

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

nPatients	T1	T2	T3	T4
ALL	6	11	101	4
subtype1	3	6	39	0
subtype2	1	0	21	3
subtype3	0	1	14	0
subtype4	1	2	12	1
subtype5	1	2	15	0

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

'RPPA cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.28 (Fisher's exact test), Q value = 0.55

Table S54. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	32	90
subtype1	15	33
subtype2	7	18
subtype3	6	10
subtype4	2	13
subtype5	2	16

Figure S50. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.296 (Fisher's exact test), Q value = 0.57

Table S55. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	60	2
subtype1	24	0
subtype2	13	1
subtype3	6	0
subtype4	7	1
subtype5	10	0

Figure S51. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.023 (Fisher's exact test), Q value = 0.2

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	57	66
subtype1	21	27
subtype2	18	7
subtype3	4	12
subtype4	5	11
subtype5	9	9

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

'RPPA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.676 (Fisher's exact test), Q value = 0.84

Table S57. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	82	28
subtype1	31	11
subtype2	16	8
subtype3	11	3
subtype4	10	4
subtype5	14	2

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

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.24 (Fisher's exact test), Q value = 0.55

Table S58. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA
ALL	108	12	2
subtype1	42	5	1
subtype2	22	2	0
subtype3	15	1	0
subtype4	11	4	1
subtype5	18	0	0

Figure S54. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RPPA cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.963 (Kruskal-Wallis (anova)), Q value = 1

Table S59. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	39	26.0 (16.7)
subtype1	18	27.8 (14.7)
subtype2	8	25.1 (21.3)
subtype3	6	24.7 (22.1)
subtype4	4	21.5 (16.5)
subtype5	3	26.7 (11.5)

Figure S55. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.65 (Kruskal-Wallis (anova)), Q value = 0.84

Table S60. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	32	1970.9 (13.4)
subtype1	13	1971.2 (13.1)
subtype2	8	1964.8 (11.8)
subtype3	5	1974.8 (15.8)
subtype4	3	1977.7 (4.0)
subtype5	3	1972.7 (22.2)

Figure S56. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RPPA cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.434 (Fisher's exact test), Q value = 0.68

Table S61. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	72	39	2	2
subtype1	25	18	1	1
subtype2	13	8	0	0
subtype3	13	2	0	1
subtype4	12	4	0	0
subtype5	9	7	1	0

Figure S57. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RPPA cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.152 (Kruskal-Wallis (anova)), Q value = 0.54

Table S62. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	122	2.9 (3.2)
subtype1	48	2.0 (2.0)
subtype2	25	3.3 (3.7)
subtype3	16	2.5 (2.9)
subtype4	15	4.3 (4.6)
subtype5	18	4.1 (3.7)

Figure S58. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.844 (Fisher's exact test), Q value = 0.94

Table S63. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	2	4	113
subtype1	1	3	44
subtype2	1	0	21
subtype3	0	0	16
subtype4	0	0	16
subtype5	0	1	16

Figure S59. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

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

Table S64. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	95
subtype1	2	36
subtype2	0	17
subtype3	0	12
subtype4	0	14
subtype5	1	16

Figure S60. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #5: 'RNAseq CNMF subtypes'

Table S65. Description of clustering approach #5: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	83	49	46

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.178 (logrank test), Q value = 0.55

Table S66. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	176	89	0.0 - 90.1 (15.0)
subtype1	82	46	0.1 - 71.7 (14.2)
subtype2	49	24	0.0 - 75.1 (15.7)
subtype3	45	19	0.3 - 90.1 (14.9)

Figure S61. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.357 (Kruskal-Wallis (anova)), Q value = 0.6

Table S67. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	178	64.6 (10.9)
subtype1	83	65.0 (10.8)
subtype2	49	65.7 (10.7)
subtype3	46	62.6 (11.5)

Figure S62. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.00178 (Fisher's exact test), Q value = 0.044

Table S68. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	119	4	4
subtype1	0	2	2	19	55	2	3
subtype2	0	1	3	6	39	0	0
subtype3	1	2	10	3	25	2	1

Figure S63. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00013 (Fisher's exact test), Q value = 0.0065

Table S69. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	142	3
subtype1	2	3	77	1
subtype2	1	10	38	0
subtype3	4	11	27	2

Figure S64. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.206 (Fisher's exact test), Q value = 0.55

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

nPatients	0	1
ALL	49	124
subtype1	23	58
subtype2	10	39
subtype3	16	27

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

'RNAseq CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.27 (Fisher's exact test), Q value = 0.55

Table S71. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	80	4
subtype1	32	3
subtype2	30	0
subtype3	18	1

Figure S66. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.674 (Fisher's exact test), Q value = 0.84

Table S72. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	98
subtype1	37	46
subtype2	20	29
subtype3	23	23

Figure S67. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.329 (Fisher's exact test), Q value = 0.59

Table S73. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	116	42
subtype1	56	16
subtype2	30	16
subtype3	30	10

Figure S68. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.108 (Fisher's exact test), Q value = 0.48

Table S74. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	147	25	4	1
subtype1	69	10	3	0
subtype2	44	4	1	0
subtype3	34	11	0	1

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

'RNAseq CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.218 (Kruskal-Wallis (anova)), Q value = 0.55

Table S75. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	57	26.8 (17.9)
subtype1	29	29.7 (15.3)
subtype2	18	23.5 (19.6)
subtype3	10	24.6 (22.2)

Figure S70. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.767 (Kruskal-Wallis (anova)), Q value = 0.9

Table S76. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	23	1969.1 (13.0)
subtype2	13	1973.1 (13.5)
subtype3	11	1971.4 (14.0)

Figure S71. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RNAseq CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.243 (Fisher's exact test), Q value = 0.55

Table S77. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	106	53	5	4
subtype1	45	25	4	3
subtype2	27	18	0	1
subtype3	34	10	1	0

Figure S72. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RNAseq CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.69 (Kruskal-Wallis (anova)), Q value = 0.84

Table S78. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	174	3.0 (3.4)
subtype1	81	2.8 (3.1)
subtype2	49	3.1 (3.4)
subtype3	44	3.2 (4.2)

Figure S73. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.13 (Fisher's exact test), Q value = 0.54

Table S79. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	6	157
subtype1	9	3	70
subtype2	2	2	44
subtype3	0	1	43

Figure S74. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.853 (Fisher's exact test), Q value = 0.94

Table S80. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	131
subtype1	2	62
subtype2	1	33
subtype3	2	36

Figure S75. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S81. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	75	75	28

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0106 (logrank test), Q value = 0.099

Table S82. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	176	89	0.0 - 90.1 (15.0)
subtype1	74	40	0.0 - 75.1 (14.1)
subtype2	75	40	0.2 - 71.7 (14.1)
subtype3	27	9	0.3 - 90.1 (21.9)

Figure S76. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.174 (Kruskal-Wallis (anova)), Q value = 0.55

Table S83. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	178	64.6 (10.9)
subtype1	75	64.9 (11.0)
subtype2	75	65.7 (10.6)
subtype3	28	60.8 (11.3)

Figure S77. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 5e-05 (Fisher's exact test), Q value = 0.0065

Table S84. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	119	4	4
subtype1	0	1	3	9	61	1	0
subtype2	0	3	3	17	47	2	3
subtype3	1	1	9	2	11	1	1

Figure S78. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 1e-04 (Fisher's exact test), Q value = 0.0065

Table S85. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	142	3
subtype1	1	12	61	1
subtype2	3	3	68	1
subtype3	3	9	13	1

Figure S79. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.00315 (Fisher's exact test), Q value = 0.059

Table S86. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	49	124
subtype1	13	62
subtype2	23	50
subtype3	13	12

Figure S80. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.103 (Fisher's exact test), Q value = 0.48

Table S87. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	80	4
subtype1	39	0
subtype2	31	3
subtype3	10	1

Figure S81. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.979 (Fisher's exact test), Q value = 1

Table S88. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	98
subtype1	34	41
subtype2	33	42
subtype3	13	15

Figure S82. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.67 (Fisher's exact test), Q value = 0.84

Table S89. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	116	42
subtype1	46	19
subtype2	53	16
subtype3	17	7

Figure S83. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.0087 (Fisher's exact test), Q value = 0.093

Table S90. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	147	25	4	1
subtype1	68	6	1	0
subtype2	61	9	3	1
subtype3	18	10	0	0

Figure S84. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.205 (Kruskal-Wallis (anova)), Q value = 0.55

Table S91. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	57	26.8 (17.9)
subtype1	31	22.9 (17.5)
subtype2	22	32.1 (18.3)
subtype3	4	28.6 (15.0)

Figure S85. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.908 (Kruskal-Wallis (anova)), Q value = 0.97

Table S92. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	26	1971.5 (12.8)
subtype2	16	1969.8 (13.7)
subtype3	5	1970.0 (16.0)

Figure S86. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RNAseq cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.146 (Fisher's exact test), Q value = 0.54

Table S93. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	106	53	5	4
subtype1	42	27	0	2
subtype2	42	21	4	2
subtype3	22	5	1	0

Figure S87. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RNAseq cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0856 (Kruskal-Wallis (anova)), Q value = 0.46

Table S94. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	174	3.0 (3.4)
subtype1	75	3.3 (3.5)
subtype2	73	2.8 (3.2)
subtype3	26	2.5 (4.2)

Figure S88. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.0989 (Fisher's exact test), Q value = 0.48

Table S95. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	6	157
subtype1	2	3	69
subtype2	9	2	62
subtype3	0	1	26

Figure S89. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.399 (Fisher's exact test), Q value = 0.64

Table S96. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	131
subtype1	2	56
subtype2	1	53
subtype3	2	22

Figure S90. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S97. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3	4
Number of samples	69	41	35	33

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.19 (logrank test), Q value = 0.55

Table S98. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	176	89	0.0 - 90.1 (15.0)
subtype1	68	38	0.2 - 66.9 (15.4)
subtype2	41	18	0.0 - 75.1 (15.3)
subtype3	35	20	1.1 - 71.7 (12.0)
subtype4	32	13	0.3 - 90.1 (15.4)

Figure S91. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.566 (Kruskal-Wallis (anova)), Q value = 0.8

Table S99. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	178	64.6 (10.9)
subtype1	69	65.0 (11.4)
subtype2	41	65.4 (9.6)
subtype3	35	65.3 (10.7)
subtype4	33	62.1 (11.9)

Figure S92. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.0557 (Fisher's exact test), Q value = 0.36

Table S100. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	119	4	4
subtype1	0	1	3	15	45	2	3
subtype2	0	1	2	7	30	1	0
subtype3	0	1	2	4	27	1	0
subtype4	1	2	8	2	17	0	1

Figure S93. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.0463 (Fisher's exact test), Q value = 0.32

Table S101. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	142	3
subtype1	1	6	61	1
subtype2	1	7	32	1
subtype3	1	3	30	1
subtype4	4	8	19	0

Figure S94. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.353 (Fisher's exact test), Q value = 0.6

Table S102. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	49	124
subtype1	20	48
subtype2	10	31
subtype3	7	27
subtype4	12	18

Figure S95. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.188 (Fisher's exact test), Q value = 0.55

Table S103. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	80	4
subtype1	25	3
subtype2	25	0
subtype3	18	0
subtype4	12	1

Figure S96. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.695 (Fisher's exact test), Q value = 0.84

Table S104. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	98
subtype1	29	40
subtype2	20	21
subtype3	18	17
subtype4	13	20

Figure S97. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY'

P value = 0.183 (Fisher's exact test), Q value = 0.55

Table S105. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	116	42
subtype1	43	20
subtype2	26	10
subtype3	25	3
subtype4	22	9

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

'MIRSEQ CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.00652 (Fisher's exact test), Q value = 0.077

Table S106. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	147	25	4	1
subtype1	56	11	1	0
subtype2	39	1	1	0
subtype3	29	3	2	1
subtype4	23	10	0	0

Figure S99. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.543 (Kruskal-Wallis (anova)), Q value = 0.78

Table S107. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	57	26.8 (17.9)
subtype1	25	29.2 (15.4)
subtype2	15	23.5 (21.7)
subtype3	11	24.6 (14.2)
subtype4	6	29.6 (25.8)

Figure S100. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CNMF' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.218 (Kruskal-Wallis (anova)), Q value = 0.55

Table S108. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	19	1967.4 (11.3)
subtype2	12	1973.9 (13.8)
subtype3	10	1968.0 (14.1)
subtype4	6	1979.5 (13.7)

Figure S101. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRSEQ CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.619 (Fisher's exact test), Q value = 0.83

Table S109. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	106	53	5	4
subtype1	37	22	3	2
subtype2	23	15	0	0
subtype3	22	10	1	1
subtype4	24	6	1	1

Figure S102. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.688 (Kruskal-Wallis (anova)), Q value = 0.84

Table S110. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	174	3.0 (3.4)
subtype1	68	2.7 (3.1)
subtype2	41	3.4 (3.7)
subtype3	34	2.8 (2.7)
subtype4	31	3.2 (4.5)

Figure S103. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.372 (Fisher's exact test), Q value = 0.61

Table S111. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	6	157
subtype1	8	2	58
subtype2	2	2	36
subtype3	1	1	32
subtype4	0	1	31

Figure S104. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.869 (Fisher's exact test), Q value = 0.94

Table S112. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	131
subtype1	1	48
subtype2	1	30
subtype3	2	27
subtype4	1	26

Figure S105. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S113. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	102	43	33

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.11 (logrank test), Q value = 0.48

Table S114. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	176	89	0.0 - 90.1 (15.0)
subtype1	101	56	0.1 - 71.7 (15.1)
subtype2	43	20	0.0 - 75.1 (12.5)
subtype3	32	13	0.3 - 90.1 (16.2)

Figure S106. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.577 (Kruskal-Wallis (anova)), Q value = 0.8

Table S115. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	178	64.6 (10.9)
subtype1	102	64.6 (11.2)
subtype2	43	65.9 (8.8)
subtype3	33	62.7 (12.6)

Figure S107. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGIC_STAGE'

P value = 0.0067 (Fisher's exact test), Q value = 0.077

Table S116. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	119	4	4
subtype1	0	2	4	21	70	2	3
subtype2	0	1	3	5	33	1	0
subtype3	1	2	8	2	16	1	1

Figure S108. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.00075 (Fisher's exact test), Q value = 0.023

Table S117. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	142	3
subtype1	2	7	92	1
subtype2	1	9	32	1
subtype3	4	8	18	1

Figure S109. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 0.111 (Fisher's exact test), Q value = 0.48

Table S118. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	49	124
subtype1	27	73
subtype2	9	34
subtype3	13	17

Figure S110. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.339 (Fisher's exact test), Q value = 0.6

Table S119. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	80	4
subtype1	45	3
subtype2	25	0
subtype3	10	1

Figure S111. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.315 (Fisher's exact test), Q value = 0.58

Table S120. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	98
subtype1	41	61
subtype2	21	22
subtype3	18	15

Figure S112. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY'

P value = 0.64 (Fisher's exact test), Q value = 0.84

Table S121. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	116	42
subtype1	68	22
subtype2	27	13
subtype3	21	7

Figure S113. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL_TYPE'

P value = 0.0423 (Fisher's exact test), Q value = 0.32

Table S122. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	147	25	4	1
subtype1	84	13	3	1
subtype2	40	2	1	0
subtype3	23	10	0	0

Figure S114. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.23 (Kruskal-Wallis (anova)), Q value = 0.55

Table S123. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	57	26.8 (17.9)
subtype1	37	29.0 (16.3)
subtype2	15	21.9 (21.4)
subtype3	5	26.0 (19.5)

Figure S115. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CHIERARCHICAL' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.827 (Kruskal-Wallis (anova)), Q value = 0.93

Table S124. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	29	1970.4 (13.2)
subtype2	12	1969.9 (14.3)
subtype3	6	1973.8 (12.4)

Figure S116. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRSEQ CHIERARCHICAL' versus 'RESIDUAL_TUMOR'

P value = 0.755 (Fisher's exact test), Q value = 0.89

Table S125. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	106	53	5	4
subtype1	57	32	4	3
subtype2	27	13	0	0
subtype3	22	8	1	1

Figure S117. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.451 (Kruskal-Wallis (anova)), Q value = 0.7

Table S126. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	174	3.0 (3.4)
subtype1	100	2.8 (3.1)
subtype2	43	3.4 (3.7)
subtype3	31	3.0 (4.2)

Figure S118. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.418 (Fisher's exact test), Q value = 0.67

Table S127. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	11	6	157
subtype1	9	3	88
subtype2	2	2	38
subtype3	0	1	31

Figure S119. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.608 (Fisher's exact test), Q value = 0.83

Table S128. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	131
subtype1	2	72
subtype2	1	32
subtype3	2	27

Figure S120. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S129. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	69	68	37

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

P value = 0.073 (logrank test), Q value = 0.44

Table S130. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	172	87	0.0 - 90.1 (15.0)
subtype1	68	40	0.0 - 66.9 (13.3)
subtype2	67	32	0.3 - 75.1 (14.9)
subtype3	37	15	0.2 - 90.1 (16.1)

Figure S121. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.519 (Kruskal-Wallis (anova)), Q value = 0.76

Table S131. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	174	64.4 (10.9)
subtype1	69	64.7 (10.0)
subtype2	68	63.5 (11.2)
subtype3	37	65.2 (12.2)

Figure S122. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0791 (Fisher's exact test), Q value = 0.44

Table S132. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	115	4	4
subtype1	0	1	2	17	44	2	3
subtype2	0	3	8	6	47	2	1
subtype3	1	1	5	5	24	0	0

Figure S123. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.146 (Fisher's exact test), Q value = 0.54

Table S133. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	138	3
subtype1	1	5	62	1
subtype2	4	12	49	2
subtype3	2	7	27	0

Figure S124. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.883 (Fisher's exact test), Q value = 0.95

Table S134. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	49	120
subtype1	21	47
subtype2	18	49
subtype3	10	24

Figure S125. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.469 (Fisher's exact test), Q value = 0.7

Table S135. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	77	4
subtype1	30	3
subtype2	37	1
subtype3	10	0

Figure S126. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.301 (Fisher's exact test), Q value = 0.57

Table S136. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	94
subtype1	35	34
subtype2	32	36
subtype3	13	24

Figure S127. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.232 (Fisher's exact test), Q value = 0.55

Table S137. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	113	41
subtype1	48	14
subtype2	44	14
subtype3	21	13

Figure S128. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.328 (Fisher's exact test), Q value = 0.59

Table S138. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	144	24	4	1
subtype1	59	7	1	1
subtype2	58	8	2	0
subtype3	27	9	1	0

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

'MIRseq Mature CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.153 (Kruskal-Wallis (anova)), Q value = 0.54

Table S139. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	27.0 (18.2)
subtype1	24	27.7 (13.7)
subtype2	18	30.7 (20.3)
subtype3	13	20.5 (21.8)

Figure S130. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'MIRseq Mature CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.986 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	20	1970.8 (11.1)
subtype2	16	1971.1 (15.3)
subtype3	11	1970.2 (14.5)

Figure S131. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRseq Mature CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.679 (Fisher's exact test), Q value = 0.84

Table S141. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	104	51	5	4
subtype1	38	22	3	1
subtype2	43	17	2	1
subtype3	23	12	0	2

Figure S132. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.729 (Kruskal-Wallis (anova)), Q value = 0.87

Table S142. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	170	2.9 (3.4)
subtype1	68	2.9 (3.2)
subtype2	67	3.1 (3.7)
subtype3	35	2.5 (3.1)

Figure S133. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.276 (Fisher's exact test), Q value = 0.55

Table S143. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	10	6	154
subtype1	6	1	61
subtype2	4	3	58
subtype3	0	2	35

Figure S134. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.842 (Fisher's exact test), Q value = 0.94

Table S144. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	130
subtype1	1	51
subtype2	3	50
subtype3	1	29

Figure S135. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S145. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	86	48	40

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

P value = 0.0458 (logrank test), Q value = 0.32

Table S146. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	172	87	0.0 - 90.1 (15.0)
subtype1	86	48	0.2 - 71.7 (14.0)
subtype2	47	24	0.0 - 42.3 (15.3)
subtype3	39	15	0.3 - 90.1 (16.0)

Figure S136. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.27 (Kruskal-Wallis (anova)), Q value = 0.55

Table S147. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	174	64.4 (10.9)
subtype1	86	64.5 (11.6)
subtype2	48	66.1 (9.5)
subtype3	40	62.1 (10.9)

Figure S137. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.00367 (Fisher's exact test), Q value = 0.061

Table S148. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	5	15	28	115	4	4
subtype1	0	2	3	19	57	2	3
subtype2	0	1	2	7	36	1	1
subtype3	1	2	10	2	22	1	0

Figure S138. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0029 (Fisher's exact test), Q value = 0.059

Table S149. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	7	24	138	3
subtype1	2	6	77	1
subtype2	1	7	39	1
subtype3	4	11	22	1

Figure S139. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.205 (Fisher's exact test), Q value = 0.55

Table S150. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	49	120
subtype1	23	61
subtype2	11	37
subtype3	15	22

Figure S140. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.669 (Fisher's exact test), Q value = 0.84

Table S151. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	77	4
subtype1	36	3
subtype2	25	1
subtype3	16	0

Figure S141. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.516 (Fisher's exact test), Q value = 0.76

Table S152. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	80	94
subtype1	36	50
subtype2	25	23
subtype3	19	21

Figure S142. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.639 (Fisher's exact test), Q value = 0.84

Table S153. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	113	41
subtype1	60	19
subtype2	27	13
subtype3	26	9

Figure S143. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.218 (Fisher's exact test), Q value = 0.55

Table S154. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	144	24	4	1
subtype1	71	10	3	1
subtype2	43	4	1	0
subtype3	30	10	0	0

Figure S144. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.253 (Kruskal-Wallis (anova)), Q value = 0.55

Table S155. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	27.0 (18.2)
subtype1	29	29.5 (17.4)
subtype2	19	22.4 (20.5)
subtype3	7	29.2 (14.6)

Figure S145. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'NUMBER_PACK_YEARS_SMOKED'

'MIRseq Mature cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.58 (Kruskal-Wallis (anova)), Q value = 0.8

Table S156. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	47	1970.7 (13.2)
subtype1	22	1970.5 (13.4)
subtype2	16	1973.0 (12.6)
subtype3	9	1967.2 (14.2)

Figure S146. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRseq Mature cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.46 (Fisher's exact test), Q value = 0.7

Table S157. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	104	51	5	4
subtype1	52	22	4	2
subtype2	25	19	0	1
subtype3	27	10	1	1

Figure S147. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.677 (Kruskal-Wallis (anova)), Q value = 0.84

Table S158. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	170	2.9 (3.4)
subtype1	84	2.6 (2.9)
subtype2	48	3.2 (3.6)
subtype3	38	3.2 (4.1)

Figure S148. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.231 (Fisher's exact test), Q value = 0.55

Table S159. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	10	6	154
subtype1	8	2	74
subtype2	2	2	43
subtype3	0	2	37

Figure S149. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.723 (Fisher's exact test), Q value = 0.87

Table S160. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	130
subtype1	2	58
subtype2	1	38
subtype3	2	34

Figure S150. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/PAAD-TP/20139891/PAAD-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/PAAD-TP/19775410/PAAD-TP.merged_data.txt
Number of patients = 185
Number of clustering approaches = 10
Number of selected clinical features = 15
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.

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