Correlation between aggregated molecular cancer subtypes and selected clinical features

Pancreatic 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/C1RR1XBH

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 14 clinical features across 174 patients, 27 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 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_T_STAGE', 'GENDER', and 'HISTOLOGICAL_TYPE'.
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 'HISTOLOGICAL_TYPE'.
Consensus hierarchical clustering analysis on RPPA data identified 7 subtypes that correlate to 'Time to Death', 'NEOPLASM_DISEASESTAGE', and 'PATHOLOGY_N_STAGE'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'NEOPLASM_DISEASESTAGE' and 'PATHOLOGY_T_STAGE'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time to Death', 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', and 'HISTOLOGICAL_TYPE'.
4 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'NEOPLASM_DISEASESTAGE', 'PATHOLOGY_T_STAGE', and 'HISTOLOGICAL_TYPE'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'Time to Death', 'NEOPLASM_DISEASESTAGE', 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 'Time to Death', 'NEOPLASM_DISEASESTAGE', 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 14 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 27 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	0.118 (0.424)	0.00851 (0.0917)	0.208 (0.54)	0.0196 (0.137)	0.0521 (0.251)	0.0036 (0.072)	0.0894 (0.348)	0.043 (0.232)	0.0604 (0.273)	0.0178 (0.137)
YEARS TO BIRTH	Kruskal-Wallis (anova)	0.964 (1.00)	0.387 (0.669)	0.591 (0.866)	0.95 (1.00)	0.424 (0.706)	0.254 (0.559)	0.561 (0.845)	0.648 (0.898)	0.662 (0.906)	0.334 (0.649)
NEOPLASM DISEASESTAGE	Fisher's exact test	0.0288 (0.183)	0.189 (0.512)	0.359 (0.658)	0.0158 (0.137)	0.00607 (0.0773)	0.00014 (0.00653)	0.0316 (0.191)	0.0142 (0.137)	0.273 (0.573)	0.00676 (0.0789)
PATHOLOGY T STAGE	Fisher's exact test	0.0261 (0.174)	0.00045 (0.0157)	0.165 (0.493)	0.118 (0.424)	0.00014 (0.00653)	8e-05 (0.00653)	0.0184 (0.137)	0.00086 (0.0241)	0.147 (0.477)	0.00248 (0.0579)
PATHOLOGY N STAGE	Fisher's exact test	0.942 (1.00)	0.222 (0.555)	0.351 (0.658)	0.0342 (0.192)	0.234 (0.559)	0.00434 (0.0759)	0.251 (0.559)	0.127 (0.432)	0.974 (1.00)	0.203 (0.537)
PATHOLOGY M STAGE	Fisher's exact test	1 (1.00)	0.407 (0.687)	0.368 (0.658)	0.689 (0.918)	0.445 (0.725)	0.161 (0.491)	0.339 (0.65)	0.23 (0.559)	0.722 (0.928)	1 (1.00)
GENDER	Fisher's exact test	0.0328 (0.191)	0.00602 (0.0773)	0.0557 (0.26)	0.0778 (0.33)	0.719 (0.928)	1 (1.00)	0.772 (0.947)	0.392 (0.669)	0.371 (0.658)	0.524 (0.806)
HISTOLOGICAL TYPE	Fisher's exact test	0.0461 (0.239)	0.099 (0.374)	0.0194 (0.137)	0.276 (0.573)	0.155 (0.483)	0.0155 (0.137)	0.00557 (0.0773)	0.0821 (0.338)	0.243 (0.559)	0.353 (0.658)
NUMBER PACK YEARS SMOKED	Kruskal-Wallis (anova)	0.297 (0.595)	0.81 (0.947)	0.19 (0.512)	0.754 (0.947)	0.326 (0.642)	0.256 (0.559)	0.673 (0.906)	0.281 (0.573)	0.179 (0.51)	0.248 (0.559)
YEAR OF TOBACCO SMOKING ONSET	Kruskal-Wallis (anova)	0.28 (0.573)	0.804 (0.947)	0.19 (0.512)	0.623 (0.881)	0.774 (0.947)	0.871 (0.982)	0.219 (0.555)	0.867 (0.982)	0.978 (1.00)	0.495 (0.786)
COMPLETENESS OF RESECTION	Fisher's exact test	0.0706 (0.309)	0.459 (0.738)	0.143 (0.476)	0.371 (0.658)	0.243 (0.559)	0.171 (0.5)	0.378 (0.661)	0.553 (0.842)	0.708 (0.928)	0.62 (0.881)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.15 (0.477)	0.5 (0.786)	0.282 (0.573)	0.124 (0.432)	0.64 (0.896)	0.0881 (0.348)	0.513 (0.799)	0.436 (0.718)	0.812 (0.947)	0.594 (0.866)
RACE	Fisher's exact test	0.671 (0.906)	0.956 (1.00)	0.871 (0.982)	0.0519 (0.251)	0.93 (1.00)	0.968 (1.00)	0.989 (1.00)	0.965 (1.00)	0.712 (0.928)	0.84 (0.972)
ETHNICITY	Fisher's exact test	1 (1.00)	0.796 (0.947)	0.876 (0.982)	1 (1.00)	0.801 (0.947)	0.576 (0.858)	0.794 (0.947)	0.783 (0.947)	0.794 (0.947)	0.62 (0.881)

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	81	68	24

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

P value = 0.118 (logrank test), Q value = 0.42

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	171	92	0.1 - 84.1 (14.2)
subtype1	79	39	0.3 - 75.1 (14.4)
subtype2	68	41	0.1 - 66.3 (12.9)
subtype3	24	12	0.2 - 84.1 (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 = 0.964 (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	173	65.0 (11.2)
subtype1	81	65.4 (10.7)
subtype2	68	64.6 (11.9)
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 'NEOPLASM_DISEASESTAGE'

P value = 0.0288 (Fisher's exact test), Q value = 0.18

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

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	26	117	5	4
subtype1	0	4	6	11	55	3	2
subtype2	0	2	1	14	47	2	2
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: 'NEOPLASM_DISEASESTAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0261 (Fisher's exact test), Q value = 0.17

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

nPatients	T1	T2	T3	T4
ALL	8	21	139	4
subtype1	5	11	62	3
subtype2	2	3	62	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 = 0.942 (Fisher's exact test), Q value = 1

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

nPatients	0	1
ALL	45	124
subtype1	21	59
subtype2	19	49
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 = 1 (Fisher's exact test), Q value = 1

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

nPatients	0	1
ALL	80	4
subtype1	40	2
subtype2	32	2
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 = 0.0328 (Fisher's exact test), Q value = 0.19

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

nPatients	FEMALE	MALE
ALL	78	95
subtype1	41	40
subtype2	32	36
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 'HISTOLOGICAL_TYPE'

P value = 0.0461 (Fisher's exact test), Q value = 0.24

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

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	145	22	4	1
subtype1	68	9	3	1
subtype2	61	6	0	0
subtype3	16	7	1	0

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.297 (Kruskal-Wallis (anova)), Q value = 0.59

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

	nPatients	Mean (Std.Dev)
ALL	54	26.5 (17.8)
subtype1	25	24.5 (18.2)
subtype2	24	29.8 (17.8)
subtype3	5	20.2 (15.8)

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

'Copy Number Ratio CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.28 (Kruskal-Wallis (anova)), Q value = 0.57

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

	nPatients	Mean (Std.Dev)
ALL	45	1970.3 (13.2)
subtype1	22	1969.8 (14.6)
subtype2	18	1973.1 (12.2)
subtype3	5	1962.6 (7.4)

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

'Copy Number Ratio CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.0706 (Fisher's exact test), Q value = 0.31

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

nPatients	R0	R1	R2	RX
ALL	108	51	2	4
subtype1	58	17	1	1
subtype2	37	25	1	1
subtype3	13	9	0	2

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.15 (Kruskal-Wallis (anova)), Q value = 0.48

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

	nPatients	Mean (Std.Dev)
ALL	170	3.0 (3.5)
subtype1	80	3.1 (3.5)
subtype2	68	2.5 (2.8)
subtype3	22	4.7 (4.8)

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

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.671 (Fisher's exact test), Q value = 0.91

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

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	7	158
subtype1	1	5	73
subtype2	2	2	62
subtype3	0	0	23

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

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

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

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

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	135
subtype1	2	68
subtype2	1	47
subtype3	0	20

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

Clustering Approach #2: 'METHLYATION CNMF'

Table S16. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3	4
Number of samples	42	46	46	39

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.00851 (logrank test), Q value = 0.092

Table S17. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	171	91	0.1 - 84.1 (14.2)
subtype1	42	27	0.2 - 71.7 (14.6)
subtype2	46	17	0.3 - 84.1 (15.6)
subtype3	45	30	0.1 - 27.8 (13.3)
subtype4	38	17	0.3 - 49.4 (12.5)

Figure S15. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.387 (Kruskal-Wallis (anova)), Q value = 0.67

Table S18. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	173	64.8 (11.1)
subtype1	42	67.4 (10.9)
subtype2	46	64.2 (10.4)
subtype3	46	63.9 (11.7)
subtype4	39	63.7 (11.3)

Figure S16. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.189 (Fisher's exact test), Q value = 0.51

Table S19. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	26	117	5	4
subtype1	0	3	4	5	28	1	1
subtype2	1	3	7	6	26	1	1
subtype3	0	0	1	10	32	1	2
subtype4	0	0	1	5	31	2	0

Figure S17. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.00045 (Fisher's exact test), Q value = 0.016

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	139	4
subtype1	3	4	33	2
subtype2	5	12	28	0
subtype3	0	2	44	0
subtype4	0	3	34	2

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.222 (Fisher's exact test), Q value = 0.56

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	45	124
subtype1	13	29
subtype2	15	28
subtype3	11	35
subtype4	6	32

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.407 (Fisher's exact test), Q value = 0.69

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	79	4
subtype1	20	1
subtype2	18	1
subtype3	18	2
subtype4	23	0

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.00602 (Fisher's exact test), Q value = 0.077

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	77	96
subtype1	10	32
subtype2	22	24
subtype3	21	25
subtype4	24	15

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.099 (Fisher's exact test), Q value = 0.37

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	145	22	4	1
subtype1	38	4	0	0
subtype2	33	11	2	0
subtype3	39	5	1	0
subtype4	35	2	1	1

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.81 (Kruskal-Wallis (anova)), Q value = 0.95

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	26.4 (17.6)
subtype1	17	26.3 (12.9)
subtype2	10	24.6 (24.6)
subtype3	14	26.0 (17.1)
subtype4	14	28.4 (19.2)

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'METHLYATION CNMF' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.804 (Kruskal-Wallis (anova)), Q value = 0.95

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	16	1969.8 (12.2)
subtype2	7	1975.4 (16.9)
subtype3	10	1970.5 (13.3)
subtype4	13	1969.3 (13.6)

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'METHLYATION CNMF' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.459 (Fisher's exact test), Q value = 0.74

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	108	51	2	4
subtype1	23	16	0	2
subtype2	32	11	0	0
subtype3	27	13	1	2
subtype4	26	11	1	0

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'METHLYATION CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.5 (Kruskal-Wallis (anova)), Q value = 0.79

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	170	3.0 (3.5)
subtype1	42	3.1 (3.6)
subtype2	44	2.8 (3.8)
subtype3	46	3.0 (3.5)
subtype4	38	3.3 (3.1)

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 0.956 (Fisher's exact test), Q value = 1

Table S29. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	7	159
subtype1	0	2	39
subtype2	1	2	42
subtype3	1	1	44
subtype4	1	2	34

Figure S27. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.796 (Fisher's exact test), Q value = 0.95

Table S30. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	136
subtype1	1	31
subtype2	0	37
subtype3	1	37
subtype4	1	31

Figure S28. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #3: 'RPPA CNMF subtypes'

Table S31. Description of clustering approach #3: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	35	19	16	18	16

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.208 (logrank test), Q value = 0.54

Table S32. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	104	58	0.1 - 71.7 (13.8)
subtype1	35	20	0.1 - 71.7 (12.9)
subtype2	19	12	2.5 - 41.3 (15.3)
subtype3	16	5	0.3 - 45.5 (18.8)
subtype4	18	11	0.3 - 68.5 (14.2)
subtype5	16	10	3.1 - 24.1 (8.7)

Figure S29. 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 = 0.591 (Kruskal-Wallis (anova)), Q value = 0.87

Table S33. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	66.0 (11.5)
subtype1	35	66.7 (11.1)
subtype2	19	68.3 (12.9)
subtype3	16	67.0 (9.4)
subtype4	18	63.3 (11.5)
subtype5	16	63.8 (12.9)

Figure S30. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.359 (Fisher's exact test), Q value = 0.66

Table S34. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	5	7	15	70	5	2
subtype1	1	4	6	24	0	0
subtype2	1	0	2	12	3	1
subtype3	1	0	2	11	1	1
subtype4	2	2	4	10	0	0
subtype5	0	1	1	13	1	0

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.165 (Fisher's exact test), Q value = 0.49

Table S35. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	6	8	86	4
subtype1	1	5	29	0
subtype2	1	0	16	2
subtype3	1	1	13	1
subtype4	3	2	13	0
subtype5	0	0	15	1

Figure S32. 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 = 0.351 (Fisher's exact test), Q value = 0.66

Table S36. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	27	76
subtype1	11	24
subtype2	4	15
subtype3	3	13
subtype4	7	10
subtype5	2	14

Figure S33. 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 = 0.368 (Fisher's exact test), Q value = 0.66

Table S37. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	47	2
subtype1	19	0
subtype2	7	1
subtype3	7	1
subtype4	6	0
subtype5	8	0

Figure S34. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.0557 (Fisher's exact test), Q value = 0.26

Table S38. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	50	54
subtype1	14	21
subtype2	15	4
subtype3	7	9
subtype4	8	10
subtype5	6	10

Figure S35. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.0194 (Fisher's exact test), Q value = 0.14

Table S39. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA
ALL	92	9	2
subtype1	34	1	0
subtype2	14	3	1
subtype3	11	4	1
subtype4	17	1	0
subtype5	16	0	0

Figure S36. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'RPPA CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.19 (Kruskal-Wallis (anova)), Q value = 0.51

Table S40. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	34	26.7 (16.6)
subtype1	13	33.5 (11.7)
subtype2	7	27.6 (21.7)
subtype3	6	17.8 (15.3)
subtype4	5	16.4 (20.1)
subtype5	3	30.0 (10.0)

Figure S37. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.19 (Kruskal-Wallis (anova)), Q value = 0.51

Table S41. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	27	1969.6 (13.4)
subtype1	9	1965.7 (11.3)
subtype2	7	1965.4 (12.6)
subtype3	4	1980.5 (6.6)
subtype4	5	1975.2 (16.0)
subtype5	2	1966.0 (25.5)

Figure S38. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RPPA CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.143 (Fisher's exact test), Q value = 0.48

Table S42. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	61	32	1	2
subtype1	20	12	0	1
subtype2	8	7	0	0
subtype3	12	3	0	0
subtype4	15	3	0	0
subtype5	6	7	1	1

Figure S39. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RPPA CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.282 (Kruskal-Wallis (anova)), Q value = 0.57

Table S43. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	103	2.9 (3.3)
subtype1	35	2.1 (2.1)
subtype2	19	3.5 (3.9)
subtype3	16	3.4 (4.3)
subtype4	17	2.5 (3.0)
subtype5	16	4.1 (3.6)

Figure S40. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.871 (Fisher's exact test), Q value = 0.98

Table S44. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	4	95
subtype1	1	3	30
subtype2	0	0	16
subtype3	0	0	16
subtype4	0	1	17
subtype5	0	0	16

Figure S41. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.876 (Fisher's exact test), Q value = 0.98

Table S45. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	2	87
subtype1	1	26
subtype2	0	15
subtype3	0	14
subtype4	0	18
subtype5	1	14

Figure S42. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S46. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4	5	6	7
Number of samples	11	19	23	15	11	18	7

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.0196 (logrank test), Q value = 0.14

Table S47. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	104	58	0.1 - 71.7 (13.8)
subtype1	11	5	3.4 - 66.3 (19.5)
subtype2	19	11	2.5 - 41.3 (15.3)
subtype3	23	12	0.1 - 71.7 (11.9)
subtype4	15	3	0.3 - 68.5 (17.8)
subtype5	11	5	0.3 - 36.2 (7.8)
subtype6	18	17	3.6 - 43.8 (14.5)
subtype7	7	5	2.0 - 19.8 (11.0)

Figure S43. 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 = 0.95 (Kruskal-Wallis (anova)), Q value = 1

Table S48. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	66.0 (11.5)
subtype1	11	66.8 (9.7)
subtype2	19	66.8 (12.4)
subtype3	23	64.0 (13.3)
subtype4	15	66.3 (7.9)
subtype5	11	65.1 (8.9)
subtype6	18	68.1 (14.7)
subtype7	7	64.6 (8.7)

Figure S44. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.0158 (Fisher's exact test), Q value = 0.14

Table S49. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	5	7	15	70	5	2
subtype1	1	0	1	9	0	0
subtype2	0	0	2	12	4	1
subtype3	3	4	3	13	0	0
subtype4	1	1	1	10	1	1
subtype5	0	2	3	6	0	0
subtype6	0	0	1	17	0	0
subtype7	0	0	4	3	0	0

Figure S45. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.118 (Fisher's exact test), Q value = 0.42

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

nPatients	T1	T2	T3	T4
ALL	6	8	86	4
subtype1	1	0	10	0
subtype2	0	0	16	3
subtype3	4	4	15	0
subtype4	1	2	11	1
subtype5	0	1	10	0
subtype6	0	1	17	0
subtype7	0	0	7	0

Figure S46. 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 = 0.0342 (Fisher's exact test), Q value = 0.19

Table S51. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	27	76
subtype1	2	9
subtype2	4	15
subtype3	10	13
subtype4	2	12
subtype5	4	7
subtype6	1	17
subtype7	4	3

Figure S47. 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 = 0.689 (Fisher's exact test), Q value = 0.92

Table S52. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	47	2
subtype1	3	0
subtype2	8	1
subtype3	9	0
subtype4	7	1
subtype5	3	0
subtype6	11	0
subtype7	6	0

Figure S48. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.0778 (Fisher's exact test), Q value = 0.33

Table S53. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	50	54
subtype1	7	4
subtype2	14	5
subtype3	9	14
subtype4	5	10
subtype5	3	8
subtype6	10	8
subtype7	2	5

Figure S49. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.276 (Fisher's exact test), Q value = 0.57

Table S54. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA
ALL	92	9	2
subtype1	11	0	0
subtype2	16	2	0
subtype3	21	1	1
subtype4	10	4	1
subtype5	11	0	0
subtype6	17	1	0
subtype7	6	1	0

Figure S50. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'RPPA cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.754 (Kruskal-Wallis (anova)), Q value = 0.95

Table S55. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	34	26.7 (16.6)
subtype1	5	33.4 (9.2)
subtype2	7	27.6 (21.7)
subtype3	8	23.2 (14.9)
subtype4	4	21.5 (16.5)
subtype5	3	28.3 (22.5)
subtype6	4	20.2 (15.9)
subtype7	3	36.7 (20.8)

Figure S51. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.623 (Kruskal-Wallis (anova)), Q value = 0.88

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	27	1969.6 (13.4)
subtype1	3	1965.7 (2.5)
subtype2	7	1965.4 (12.6)
subtype3	7	1968.7 (15.6)
subtype4	3	1977.7 (4.0)
subtype5	3	1975.3 (11.7)
subtype6	2	1969.5 (30.4)
subtype7	2	1972.5 (23.3)

Figure S52. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RPPA cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.371 (Fisher's exact test), Q value = 0.66

Table S57. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	61	32	1	2
subtype1	5	6	0	0
subtype2	8	7	0	0
subtype3	12	8	0	1
subtype4	11	4	0	0
subtype5	9	1	0	1
subtype6	10	5	1	0
subtype7	6	1	0	0

Figure S53. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RPPA cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.124 (Kruskal-Wallis (anova)), Q value = 0.43

Table S58. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	103	2.9 (3.3)
subtype1	11	2.3 (2.1)
subtype2	19	3.7 (4.0)
subtype3	23	1.9 (2.3)
subtype4	14	4.4 (4.8)
subtype5	11	2.5 (2.8)
subtype6	18	3.7 (3.1)
subtype7	7	1.1 (1.5)

Figure S54. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.0519 (Fisher's exact test), Q value = 0.25

Table S59. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	4	95
subtype1	1	0	10
subtype2	0	0	16
subtype3	0	1	21
subtype4	0	0	15
subtype5	0	0	11
subtype6	0	1	17
subtype7	0	2	5

Figure S55. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

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

Table S60. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	2	87
subtype1	0	7
subtype2	0	15
subtype3	1	20
subtype4	0	13
subtype5	0	10
subtype6	1	16
subtype7	0	6

Figure S56. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #5: 'RNAseq CNMF subtypes'

Table S61. Description of clustering approach #5: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	75	47	45

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.0521 (logrank test), Q value = 0.25

Table S62. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	85	0.1 - 84.1 (13.8)
subtype1	74	44	0.1 - 71.7 (13.1)
subtype2	47	24	0.3 - 75.1 (15.7)
subtype3	44	17	0.3 - 84.1 (13.8)

Figure S57. 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 = 0.424 (Kruskal-Wallis (anova)), Q value = 0.71

Table S63. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	167	64.6 (11.1)
subtype1	75	65.0 (10.9)
subtype2	47	65.8 (10.9)
subtype3	45	62.8 (11.5)

Figure S58. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00607 (Fisher's exact test), Q value = 0.077

Table S64. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	114	4	3
subtype1	0	2	2	16	51	2	2
subtype2	0	1	2	6	38	0	0
subtype3	1	3	9	3	25	2	1

Figure S59. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00014 (Fisher's exact test), Q value = 0.0065

Table S65. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	134	3
subtype1	2	2	70	1
subtype2	1	9	37	0
subtype3	5	10	27	2

Figure S60. 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 = 0.234 (Fisher's exact test), Q value = 0.56

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

nPatients	0	1
ALL	44	119
subtype1	20	54
subtype2	9	38
subtype3	15	27

Figure S61. 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 = 0.445 (Fisher's exact test), Q value = 0.72

Table S67. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	75	3
subtype1	30	2
subtype2	28	0
subtype3	17	1

Figure S62. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.719 (Fisher's exact test), Q value = 0.93

Table S68. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	75	92
subtype1	34	41
subtype2	19	28
subtype3	22	23

Figure S63. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.155 (Fisher's exact test), Q value = 0.48

Table S69. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	139	22	4	1
subtype1	63	8	3	0
subtype2	42	4	1	0
subtype3	34	10	0	1

Figure S64. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'RNAseq CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.326 (Kruskal-Wallis (anova)), Q value = 0.64

Table S70. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	26.4 (17.6)
subtype1	28	28.6 (14.4)
subtype2	18	23.5 (19.6)
subtype3	9	25.6 (23.3)

Figure S65. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.774 (Kruskal-Wallis (anova)), Q value = 0.95

Table S71. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	23	1969.1 (13.0)
subtype2	13	1973.1 (13.5)
subtype3	10	1971.1 (14.7)

Figure S66. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RNAseq CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.243 (Fisher's exact test), Q value = 0.56

Table S72. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	103	51	2	4
subtype1	42	25	1	3
subtype2	27	17	0	1
subtype3	34	9	1	0

Figure S67. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RNAseq CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.64 (Kruskal-Wallis (anova)), Q value = 0.9

Table S73. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	164	3.0 (3.5)
subtype1	74	2.7 (3.1)
subtype2	47	3.2 (3.4)
subtype3	43	3.2 (4.2)

Figure S68. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.93 (Fisher's exact test), Q value = 1

Table S74. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	154
subtype1	2	3	69
subtype2	1	2	43
subtype3	0	1	42

Figure S69. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.801 (Fisher's exact test), Q value = 0.95

Table S75. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	130
subtype1	2	61
subtype2	0	33
subtype3	1	36

Figure S70. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S76. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	73	67	27

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0036 (logrank test), Q value = 0.072

Table S77. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	85	0.1 - 84.1 (13.8)
subtype1	72	39	0.1 - 75.1 (12.9)
subtype2	67	38	0.2 - 71.7 (12.9)
subtype3	26	8	0.3 - 84.1 (20.3)

Figure S71. 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 = 0.254 (Kruskal-Wallis (anova)), Q value = 0.56

Table S78. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	167	64.6 (11.1)
subtype1	73	64.9 (11.1)
subtype2	67	65.7 (10.8)
subtype3	27	61.1 (11.5)

Figure S72. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00014 (Fisher's exact test), Q value = 0.0065

Table S79. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	114	4	3
subtype1	0	1	2	9	60	1	0
subtype2	0	3	3	14	43	2	2
subtype3	1	2	8	2	11	1	1

Figure S73. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 8e-05 (Fisher's exact test), Q value = 0.0065

Table S80. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	134	3
subtype1	1	11	60	1
subtype2	3	2	61	1
subtype3	4	8	13	1

Figure S74. 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 = 0.00434 (Fisher's exact test), Q value = 0.076

Table S81. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	44	119
subtype1	12	61
subtype2	20	46
subtype3	12	12

Figure S75. 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 = 0.161 (Fisher's exact test), Q value = 0.49

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

nPatients	0	1
ALL	75	3
subtype1	37	0
subtype2	29	2
subtype3	9	1

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

'RNAseq cHierClus subtypes' versus 'GENDER'

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

Table S83. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	75	92
subtype1	33	40
subtype2	30	37
subtype3	12	15

Figure S77. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.0155 (Fisher's exact test), Q value = 0.14

Table S84. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	139	22	4	1
subtype1	66	6	1	0
subtype2	55	7	3	1
subtype3	18	9	0	0

Figure S78. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.256 (Kruskal-Wallis (anova)), Q value = 0.56

Table S85. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	26.4 (17.6)
subtype1	31	22.9 (17.5)
subtype2	21	30.7 (17.6)
subtype3	3	33.0 (14.7)

Figure S79. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.871 (Kruskal-Wallis (anova)), Q value = 0.98

Table S86. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	26	1971.5 (12.8)
subtype2	16	1969.8 (13.7)
subtype3	4	1969.0 (18.3)

Figure S80. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'RNAseq cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.171 (Fisher's exact test), Q value = 0.5

Table S87. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	103	51	2	4
subtype1	42	26	0	2
subtype2	39	21	1	2
subtype3	22	4	1	0

Figure S81. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'RNAseq cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0881 (Kruskal-Wallis (anova)), Q value = 0.35

Table S88. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	164	3.0 (3.5)
subtype1	73	3.4 (3.5)
subtype2	66	2.7 (3.2)
subtype3	25	2.6 (4.3)

Figure S82. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.968 (Fisher's exact test), Q value = 1

Table S89. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	154
subtype1	1	3	68
subtype2	2	2	61
subtype3	0	1	25

Figure S83. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.576 (Fisher's exact test), Q value = 0.86

Table S90. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	130
subtype1	1	56
subtype2	1	52
subtype3	1	22

Figure S84. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S91. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3	4
Number of samples	61	39	34	33

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0894 (logrank test), Q value = 0.35

Table S92. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	85	0.1 - 84.1 (13.8)
subtype1	60	36	0.2 - 66.3 (15.1)
subtype2	39	17	0.1 - 75.1 (14.4)
subtype3	34	19	0.7 - 71.7 (11.7)
subtype4	32	13	0.3 - 84.1 (15.4)

Figure S85. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.561 (Kruskal-Wallis (anova)), Q value = 0.84

Table S93. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	167	64.6 (11.1)
subtype1	61	64.9 (11.6)
subtype2	39	65.4 (9.8)
subtype3	34	65.6 (10.6)
subtype4	33	62.1 (11.9)

Figure S86. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.0316 (Fisher's exact test), Q value = 0.19

Table S94. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	114	4	3
subtype1	0	1	3	12	41	2	2
subtype2	0	1	1	7	29	1	0
subtype3	0	1	1	4	27	1	0
subtype4	1	3	8	2	17	0	1

Figure S87. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.0184 (Fisher's exact test), Q value = 0.14

Table S95. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	134	3
subtype1	1	5	54	1
subtype2	1	6	31	1
subtype3	1	2	30	1
subtype4	5	8	19	0

Figure S88. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.251 (Fisher's exact test), Q value = 0.56

Table S96. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	44	119
subtype1	17	44
subtype2	9	30
subtype3	6	27
subtype4	12	18

Figure S89. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.339 (Fisher's exact test), Q value = 0.65

Table S97. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	75	3
subtype1	23	2
subtype2	23	0
subtype3	17	0
subtype4	12	1

Figure S90. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.772 (Fisher's exact test), Q value = 0.95

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

nPatients	FEMALE	MALE
ALL	75	92
subtype1	26	35
subtype2	19	20
subtype3	17	17
subtype4	13	20

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

'MIRSEQ CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 0.00557 (Fisher's exact test), Q value = 0.077

Table S99. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	139	22	4	1
subtype1	50	9	1	0
subtype2	37	1	1	0
subtype3	29	2	2	1
subtype4	23	10	0	0

Figure S92. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'MIRSEQ CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.673 (Kruskal-Wallis (anova)), Q value = 0.91

Table S100. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	26.4 (17.6)
subtype1	24	27.9 (14.3)
subtype2	15	23.5 (21.7)
subtype3	10	25.5 (14.6)
subtype4	6	29.6 (25.8)

Figure S93. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CNMF' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.219 (Kruskal-Wallis (anova)), Q value = 0.56

Table S101. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	19	1967.4 (11.3)
subtype2	12	1973.9 (13.8)
subtype3	9	1967.3 (14.8)
subtype4	6	1979.5 (13.7)

Figure S94. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRSEQ CNMF' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.378 (Fisher's exact test), Q value = 0.66

Table S102. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	103	51	2	4
subtype1	34	22	0	2
subtype2	23	14	0	0
subtype3	22	9	1	1
subtype4	24	6	1	1

Figure S95. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRSEQ CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.513 (Kruskal-Wallis (anova)), Q value = 0.8

Table S103. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	164	3.0 (3.5)
subtype1	61	2.7 (3.2)
subtype2	39	3.5 (3.7)
subtype3	33	2.9 (2.7)
subtype4	31	3.2 (4.5)

Figure S96. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.989 (Fisher's exact test), Q value = 1

Table S104. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	154
subtype1	1	2	57
subtype2	1	2	35
subtype3	1	1	31
subtype4	0	1	31

Figure S97. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.794 (Fisher's exact test), Q value = 0.95

Table S105. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	130
subtype1	1	47
subtype2	0	30
subtype3	1	27
subtype4	1	26

Figure S98. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S106. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	94	41	32

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.043 (logrank test), Q value = 0.23

Table S107. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	85	0.1 - 84.1 (13.8)
subtype1	93	54	0.1 - 71.7 (13.7)
subtype2	41	19	0.3 - 75.1 (12.4)
subtype3	31	12	0.3 - 84.1 (14.9)

Figure S99. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.648 (Kruskal-Wallis (anova)), Q value = 0.9

Table S108. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	167	64.6 (11.1)
subtype1	94	64.6 (11.4)
subtype2	41	66.0 (8.9)
subtype3	32	63.0 (12.7)

Figure S100. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.0142 (Fisher's exact test), Q value = 0.14

Table S109. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	114	4	3
subtype1	0	2	4	18	66	2	2
subtype2	0	1	2	5	32	1	0
subtype3	1	3	7	2	16	1	1

Figure S101. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.00086 (Fisher's exact test), Q value = 0.024

Table S110. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	134	3
subtype1	2	6	85	1
subtype2	1	8	31	1
subtype3	5	7	18	1

Figure S102. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 0.127 (Fisher's exact test), Q value = 0.43

Table S111. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	44	119
subtype1	24	69
subtype2	8	33
subtype3	12	17

Figure S103. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.23 (Fisher's exact test), Q value = 0.56

Table S112. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	75	3
subtype1	43	2
subtype2	23	0
subtype3	9	1

Figure S104. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.392 (Fisher's exact test), Q value = 0.67

Table S113. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	75	92
subtype1	38	56
subtype2	20	21
subtype3	17	15

Figure S105. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL_TYPE'

P value = 0.0821 (Fisher's exact test), Q value = 0.34

Table S114. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	139	22	4	1
subtype1	78	11	3	1
subtype2	38	2	1	0
subtype3	23	9	0	0

Figure S106. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.281 (Kruskal-Wallis (anova)), Q value = 0.57

Table S115. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	55	26.4 (17.6)
subtype1	36	28.1 (15.6)
subtype2	15	21.9 (21.4)
subtype3	4	28.8 (21.4)

Figure S107. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CHIERARCHICAL' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.867 (Kruskal-Wallis (anova)), Q value = 0.98

Table S116. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	29	1970.4 (13.2)
subtype2	12	1969.9 (14.3)
subtype3	5	1973.8 (13.8)

Figure S108. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRSEQ CHIERARCHICAL' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.553 (Fisher's exact test), Q value = 0.84

Table S117. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	103	51	2	4
subtype1	54	32	1	3
subtype2	27	12	0	0
subtype3	22	7	1	1

Figure S109. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.436 (Kruskal-Wallis (anova)), Q value = 0.72

Table S118. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	164	3.0 (3.5)
subtype1	93	2.8 (3.1)
subtype2	41	3.5 (3.7)
subtype3	30	3.1 (4.3)

Figure S110. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.965 (Fisher's exact test), Q value = 1

Table S119. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	154
subtype1	2	3	87
subtype2	1	2	37
subtype3	0	1	30

Figure S111. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.783 (Fisher's exact test), Q value = 0.95

Table S120. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	130
subtype1	2	71
subtype2	0	32
subtype3	1	27

Figure S112. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S121. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	63	65	36

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

P value = 0.0604 (logrank test), Q value = 0.27

Table S122. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	162	83	0.1 - 84.1 (14.0)
subtype1	62	38	0.1 - 66.3 (12.9)
subtype2	64	30	0.3 - 75.1 (14.4)
subtype3	36	15	0.2 - 84.1 (16.0)

Figure S113. 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 = 0.662 (Kruskal-Wallis (anova)), Q value = 0.91

Table S123. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	164	64.4 (11.1)
subtype1	63	64.5 (10.2)
subtype2	65	64.0 (11.3)
subtype3	36	65.2 (12.4)

Figure S114. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.273 (Fisher's exact test), Q value = 0.57

Table S124. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	111	4	3
subtype1	0	1	2	14	42	2	2
subtype2	0	4	7	6	45	2	1
subtype3	1	1	4	5	24	0	0

Figure S115. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.147 (Fisher's exact test), Q value = 0.48

Table S125. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	131	3
subtype1	1	4	57	1
subtype2	5	11	47	2
subtype3	2	6	27	0

Figure S116. 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 = 0.974 (Fisher's exact test), Q value = 1

Table S126. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	44	116
subtype1	18	45
subtype2	17	47
subtype3	9	24

Figure S117. 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 = 0.722 (Fisher's exact test), Q value = 0.93

Table S127. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	73	3
subtype1	29	2
subtype2	35	1
subtype3	9	0

Figure S118. 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 = 0.371 (Fisher's exact test), Q value = 0.66

Table S128. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	75	89
subtype1	32	31
subtype2	30	35
subtype3	13	23

Figure S119. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.243 (Fisher's exact test), Q value = 0.56

Table S129. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	136	22	4	1
subtype1	54	6	1	1
subtype2	56	7	2	0
subtype3	26	9	1	0

Figure S120. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.179 (Kruskal-Wallis (anova)), Q value = 0.51

Table S130. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	53	26.6 (17.9)
subtype1	24	27.7 (13.7)
subtype2	16	29.8 (19.9)
subtype3	13	20.5 (21.8)

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

'MIRseq Mature CNMF subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.978 (Kruskal-Wallis (anova)), Q value = 1

Table S131. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	20	1970.8 (11.1)
subtype2	15	1970.9 (15.8)
subtype3	11	1970.2 (14.5)

Figure S122. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

'MIRseq Mature CNMF subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.708 (Fisher's exact test), Q value = 0.93

Table S132. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	102	49	2	4
subtype1	36	22	1	1
subtype2	43	16	1	1
subtype3	23	11	0	2

Figure S123. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.812 (Kruskal-Wallis (anova)), Q value = 0.95

Table S133. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	161	2.9 (3.4)
subtype1	63	2.9 (3.2)
subtype2	64	3.2 (3.8)
subtype3	34	2.5 (3.1)

Figure S124. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.712 (Fisher's exact test), Q value = 0.93

Table S134. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	151
subtype1	1	1	60
subtype2	2	3	57
subtype3	0	2	34

Figure S125. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.794 (Fisher's exact test), Q value = 0.95

Table S135. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	129
subtype1	1	50
subtype2	2	50
subtype3	0	29

Figure S126. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S136. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	79	46	39

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

P value = 0.0178 (logrank test), Q value = 0.14

Table S137. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	162	83	0.1 - 84.1 (14.0)
subtype1	79	46	0.2 - 71.7 (12.9)
subtype2	45	23	0.1 - 34.1 (15.1)
subtype3	38	14	0.3 - 84.1 (14.8)

Figure S127. 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 = 0.334 (Kruskal-Wallis (anova)), Q value = 0.65

Table S138. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	164	64.4 (11.1)
subtype1	79	64.5 (11.8)
subtype2	46	66.2 (9.6)
subtype3	39	62.3 (11.0)

Figure S128. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'NEOPLASM_DISEASESTAGE'

P value = 0.00676 (Fisher's exact test), Q value = 0.079

Table S139. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE IIA	STAGE IIB	STAGE III	STAGE IV
ALL	1	6	13	25	111	4	3
subtype1	0	2	3	16	54	2	2
subtype2	0	1	1	7	35	1	1
subtype3	1	3	9	2	22	1	0

Figure S129. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'NEOPLASM_DISEASESTAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00248 (Fisher's exact test), Q value = 0.058

Table S140. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	8	21	131	3
subtype1	2	5	71	1
subtype2	1	6	38	1
subtype3	5	10	22	1

Figure S130. 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 = 0.203 (Fisher's exact test), Q value = 0.54

Table S141. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	44	116
subtype1	20	58
subtype2	10	36
subtype3	14	22

Figure S131. 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 = 1 (Fisher's exact test), Q value = 1

Table S142. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	73	3
subtype1	35	2
subtype2	23	1
subtype3	15	0

Figure S132. 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 = 0.524 (Fisher's exact test), Q value = 0.81

Table S143. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	75	89
subtype1	33	46
subtype2	24	22
subtype3	18	21

Figure S133. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.353 (Fisher's exact test), Q value = 0.66

Table S144. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

nPatients	PANCREAS-ADENOCARCINOMA DUCTAL TYPE	PANCREAS-ADENOCARCINOMA-OTHER SUBTYPE	PANCREAS-COLLOID (MUCINOUS NON-CYSTIC) CARCINOMA	PANCREAS-UNDIFFERENTIATED CARCINOMA
ALL	136	22	4	1
subtype1	65	9	3	1
subtype2	41	4	1	0
subtype3	30	9	0	0

Figure S134. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'HISTOLOGICAL_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.248 (Kruskal-Wallis (anova)), Q value = 0.56

Table S145. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	53	26.6 (17.9)
subtype1	28	28.4 (16.7)
subtype2	19	22.4 (20.5)
subtype3	6	31.5 (14.5)

Figure S135. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'NUMBER_PACK_YEARS_SMOKED'

'MIRseq Mature cHierClus subtypes' versus 'YEAR_OF_TOBACCO_SMOKING_ONSET'

P value = 0.495 (Kruskal-Wallis (anova)), Q value = 0.79

Table S146. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'YEAR_OF_TOBACCO_SMOKING_ONSET'

	nPatients	Mean (Std.Dev)
ALL	46	1970.7 (13.3)
subtype1	22	1970.5 (13.4)
subtype2	16	1973.0 (12.6)
subtype3	8	1966.4 (15.0)

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

'MIRseq Mature cHierClus subtypes' versus 'COMPLETENESS_OF_RESECTION'

P value = 0.62 (Fisher's exact test), Q value = 0.88

Table S147. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

nPatients	R0	R1	R2	RX
ALL	102	49	2	4
subtype1	50	22	1	2
subtype2	25	18	0	1
subtype3	27	9	1	1

Figure S137. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'COMPLETENESS_OF_RESECTION'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.594 (Kruskal-Wallis (anova)), Q value = 0.87

Table S148. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	161	2.9 (3.4)
subtype1	78	2.6 (2.9)
subtype2	46	3.3 (3.6)
subtype3	37	3.2 (4.1)

Figure S138. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.84 (Fisher's exact test), Q value = 0.97

Table S149. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	3	6	151
subtype1	2	2	73
subtype2	1	2	42
subtype3	0	2	36

Figure S139. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.62 (Fisher's exact test), Q value = 0.88

Table S150. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	129
subtype1	2	57
subtype2	0	38
subtype3	1	34

Figure S140. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/PAAD-TP/15107801/PAAD-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/PAAD-TP/15087151/PAAD-TP.merged_data.txt
Number of patients = 174
Number of clustering approaches = 10
Number of selected clinical features = 14
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