Correlation between aggregated molecular cancer subtypes and selected clinical features

Skin Cutaneous Melanoma (Metastatic)

15 July 2014 | analyses__2014_07_15

Maintainer Information

Citation Information

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

Cite as Broad Institute TCGA Genome Data Analysis Center (2014): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1RV0MG7

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 295 patients, 13 significant findings detected with P value < 0.05 and Q value < 0.25.

4 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes do not correlate to any clinical features.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time from Specimen Diagnosis to Death'.
CNMF clustering analysis on RPPA data identified 4 subtypes that do not correlate to any clinical features.
Consensus hierarchical clustering analysis on RPPA data identified 4 subtypes that do not correlate to any clinical features.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time from Specimen Diagnosis to Death', 'Time to Death', 'PRIMARY.SITE.OF.DISEASE', and 'BRESLOW.THICKNESS'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time from Specimen Diagnosis to Death', 'Time to Death', 'PRIMARY.SITE.OF.DISEASE', 'PATHOLOGY.T.STAGE', and 'BRESLOW.THICKNESS'.
5 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'PATHOLOGY.T.STAGE'.
4 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes do not correlate to any clinical features.
3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'PRIMARY.SITE.OF.DISEASE' and 'PATHOLOGY.T.STAGE'.
4 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes do not correlate to any clinical features.

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, 13 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 from Specimen Diagnosis to Death	logrank test	0.0908 (1.00)	0.000851 (0.107)	0.174 (1.00)	0.312 (1.00)	1.11e-05 (0.00149)	4.84e-08 (6.58e-06)	0.277 (1.00)	0.316 (1.00)	0.0392 (1.00)	0.0835 (1.00)
Time to Death	logrank test	0.114 (1.00)	0.0445 (1.00)	0.0986 (1.00)	0.762 (1.00)	0.000615 (0.0781)	8.51e-05 (0.0112)	0.465 (1.00)	0.603 (1.00)	0.0706 (1.00)	0.465 (1.00)
AGE	Kruskal-Wallis (anova)	0.0135 (1.00)	0.295 (1.00)	0.305 (1.00)	0.548 (1.00)	0.00386 (0.471)	0.0021 (0.259)	0.121 (1.00)	0.773 (1.00)	0.255 (1.00)	0.819 (1.00)
PRIMARY SITE OF DISEASE	Fisher's exact test	0.0765 (1.00)	0.00545 (0.654)	0.101 (1.00)	0.351 (1.00)	5e-05 (0.00665)	1e-05 (0.00135)	0.0089 (1.00)	0.425 (1.00)	0.00187 (0.232)	0.00419 (0.507)
NEOPLASM DISEASESTAGE	Fisher's exact test	0.637 (1.00)	0.315 (1.00)	0.559 (1.00)	0.3 (1.00)	0.172 (1.00)	0.173 (1.00)	0.0717 (1.00)	0.346 (1.00)	0.304 (1.00)	0.494 (1.00)
PATHOLOGY T STAGE	Fisher's exact test	0.557 (1.00)	0.09 (1.00)	0.46 (1.00)	0.302 (1.00)	0.0217 (1.00)	0.00036 (0.0461)	0.00179 (0.224)	0.215 (1.00)	0.00012 (0.0155)	0.134 (1.00)
PATHOLOGY N STAGE	Fisher's exact test	0.843 (1.00)	0.118 (1.00)	0.326 (1.00)	0.579 (1.00)	0.799 (1.00)	0.995 (1.00)	0.08 (1.00)	0.396 (1.00)	0.798 (1.00)	0.932 (1.00)
PATHOLOGY M STAGE	Fisher's exact test	0.548 (1.00)	0.728 (1.00)	0.383 (1.00)	0.811 (1.00)	0.945 (1.00)	0.223 (1.00)	0.612 (1.00)	0.103 (1.00)	0.988 (1.00)	0.278 (1.00)
MELANOMA ULCERATION	Fisher's exact test	0.726 (1.00)	0.107 (1.00)	0.331 (1.00)	0.843 (1.00)	0.214 (1.00)	0.12 (1.00)	0.401 (1.00)	0.539 (1.00)	0.84 (1.00)	0.902 (1.00)
MELANOMA PRIMARY KNOWN	Fisher's exact test	0.789 (1.00)	0.283 (1.00)	0.678 (1.00)	0.132 (1.00)	0.0992 (1.00)	0.342 (1.00)	0.0281 (1.00)	0.207 (1.00)	0.241 (1.00)	0.104 (1.00)
BRESLOW THICKNESS	Kruskal-Wallis (anova)	0.515 (1.00)	0.39 (1.00)	0.224 (1.00)	0.882 (1.00)	8.04e-05 (0.0106)	0.000105 (0.0137)	0.0548 (1.00)	0.202 (1.00)	0.0949 (1.00)	0.176 (1.00)
GENDER	Fisher's exact test	0.653 (1.00)	0.818 (1.00)	0.839 (1.00)	0.275 (1.00)	0.467 (1.00)	0.633 (1.00)	0.383 (1.00)	0.49 (1.00)	0.305 (1.00)	0.753 (1.00)
RACE	Fisher's exact test	0.769 (1.00)	0.852 (1.00)			0.528 (1.00)	0.739 (1.00)	0.19 (1.00)	0.358 (1.00)	0.0575 (1.00)	0.663 (1.00)
ETHNICITY	Fisher's exact test	0.665 (1.00)	1 (1.00)			0.112 (1.00)	0.21 (1.00)	0.692 (1.00)	0.669 (1.00)	0.46 (1.00)	0.877 (1.00)

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

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

Cluster Labels	1	2	3	4
Number of samples	82	53	46	100

'Copy Number Ratio CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.0908 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	270	133	0.0 - 124.3 (14.1)
subtype1	79	42	0.0 - 122.7 (12.9)
subtype2	52	28	0.1 - 111.8 (12.0)
subtype3	44	22	1.5 - 98.3 (16.8)
subtype4	95	41	0.0 - 124.3 (18.0)

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

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

P value = 0.114 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	276	134	0.2 - 357.4 (47.7)
subtype1	81	43	0.2 - 268.9 (41.6)
subtype2	52	28	6.4 - 294.8 (48.1)
subtype3	45	22	3.2 - 357.4 (58.8)
subtype4	98	41	0.2 - 314.6 (49.2)

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

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.0135 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	277	55.8 (15.5)
subtype1	81	59.2 (16.2)
subtype2	52	51.6 (15.8)
subtype3	45	51.6 (15.8)
subtype4	99	57.0 (13.8)

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

'Copy Number Ratio CNMF subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.0765 (Fisher's exact test), Q value = 1

Table S5. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	37	4	60	179
subtype1	16	0	14	51
subtype2	9	2	11	31
subtype3	6	1	13	26
subtype4	6	1	22	71

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

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

P value = 0.637 (Fisher's exact test), Q value = 1

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

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	23	10	25	17	10	14	10	32	14	23	48	12
subtype1	1	1	5	3	8	5	3	5	4	9	5	11	14	3
subtype2	2	0	4	2	7	4	0	2	4	7	2	1	10	3
subtype3	4	2	3	2	2	4	2	0	1	6	2	5	6	4
subtype4	3	3	11	3	8	4	5	7	1	10	5	6	18	2

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

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

P value = 0.557 (Fisher's exact test), Q value = 1

Table S7. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	53	60	55	57
subtype1	17	18	18	19
subtype2	10	13	6	14
subtype3	12	9	8	6
subtype4	14	20	23	18

Figure S6. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.843 (Fisher's exact test), Q value = 1

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

nPatients	N0	N1	N2	N3
ALL	141	51	32	37
subtype1	42	16	13	9
subtype2	28	10	3	7
subtype3	24	10	6	5
subtype4	47	15	10	16

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

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

P value = 0.548 (Fisher's exact test), Q value = 1

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

nPatients	M0	M1	M1A	M1B	M1C
ALL	252	4	2	2	5
subtype1	77	1	0	0	2
subtype2	46	1	1	0	1
subtype3	42	1	0	1	2
subtype4	87	1	1	1	0

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

'Copy Number Ratio CNMF subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.726 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	104	72
subtype1	32	25
subtype2	17	15
subtype3	16	10
subtype4	39	22

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

'Copy Number Ratio CNMF subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.789 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	34	247
subtype1	10	72
subtype2	8	45
subtype3	6	40
subtype4	10	90

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

'Copy Number Ratio CNMF subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.515 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	206	3.5 (4.9)
subtype1	67	3.4 (4.2)
subtype2	37	3.3 (3.2)
subtype3	33	2.8 (3.3)
subtype4	69	4.1 (6.7)

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

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.653 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	108	173
subtype1	34	48
subtype2	23	30
subtype3	16	30
subtype4	35	65

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

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.769 (Fisher's exact test), Q value = 1

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

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	275
subtype1	1	0	81
subtype2	2	0	51
subtype3	1	0	45
subtype4	1	1	98

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 = 0.665 (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	274
subtype1	2	80
subtype2	0	52
subtype3	0	44
subtype4	1	98

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
Number of samples	104	108	73

'METHLYATION CNMF' versus 'Time from Specimen Diagnosis to Death'

P value = 0.000851 (logrank test), Q value = 0.11

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	134	0.0 - 124.3 (14.1)
subtype1	99	51	0.1 - 122.7 (15.9)
subtype2	101	42	0.0 - 124.3 (18.0)
subtype3	72	41	0.0 - 111.1 (10.6)

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

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.0445 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	135	0.2 - 357.4 (47.5)
subtype1	102	52	2.6 - 357.4 (56.2)
subtype2	105	42	0.2 - 346.0 (46.3)
subtype3	72	41	0.2 - 247.0 (39.9)

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

'METHLYATION CNMF' versus 'AGE'

P value = 0.295 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	280	55.7 (15.4)
subtype1	102	57.2 (16.1)
subtype2	106	55.0 (14.7)
subtype3	72	54.7 (15.5)

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

'METHLYATION CNMF' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.00545 (Fisher's exact test), Q value = 0.65

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	39	4	61	180
subtype1	16	2	25	61
subtype2	6	1	19	81
subtype3	17	1	17	38

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

'METHLYATION CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.315 (Fisher's exact test), Q value = 1

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

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	23	10	25	17	10	14	10	32	14	24	49	12
subtype1	6	2	7	6	10	5	5	7	5	9	8	7	13	3
subtype2	3	2	12	3	8	3	3	2	1	15	4	10	22	6
subtype3	1	2	4	1	7	9	2	5	4	8	2	7	14	3

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

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

P value = 0.09 (Fisher's exact test), Q value = 1

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	53	60	57	58
subtype1	20	20	20	24
subtype2	26	18	23	17
subtype3	7	22	14	17

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.118 (Fisher's exact test), Q value = 1

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	141	52	34	37
subtype1	59	13	15	11
subtype2	43	24	12	19
subtype3	39	15	7	7

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

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

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

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	2	5
subtype1	96	1	0	1	2
subtype2	93	1	1	1	3
subtype3	65	2	1	0	0

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

'METHLYATION CNMF' versus 'MELANOMA.ULCERATION'

P value = 0.107 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	104	74
subtype1	43	30
subtype2	43	22
subtype3	18	22

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

'METHLYATION CNMF' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.283 (Fisher's exact test), Q value = 1

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	34	251
subtype1	9	95
subtype2	17	91
subtype3	8	65

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

'METHLYATION CNMF' versus 'BRESLOW.THICKNESS'

P value = 0.39 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	208	3.6 (4.9)
subtype1	79	3.4 (2.9)
subtype2	73	3.8 (6.6)
subtype3	56	3.6 (4.8)

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

'METHLYATION CNMF' versus 'GENDER'

P value = 0.818 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	110	175
subtype1	38	66
subtype2	44	64
subtype3	28	45

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

'METHLYATION CNMF' versus 'RACE'

P value = 0.852 (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	5	1	277
subtype1	2	0	102
subtype2	1	1	104
subtype3	2	0	71

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

'METHLYATION CNMF' versus 'ETHNICITY'

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

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

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	276
subtype1	1	100
subtype2	1	105
subtype3	1	71

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
Number of samples	47	58	26	38

'RPPA CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.174 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	161	89	0.2 - 124.3 (14.8)
subtype1	45	22	3.5 - 111.9 (14.6)
subtype2	55	31	0.2 - 114.2 (20.4)
subtype3	25	15	0.2 - 81.2 (8.7)
subtype4	36	21	1.1 - 124.3 (20.5)

Figure S29. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.0986 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	90	0.2 - 357.4 (50.9)
subtype1	45	22	9.9 - 346.0 (61.0)
subtype2	58	32	0.2 - 357.4 (55.2)
subtype3	26	15	0.2 - 162.1 (46.1)
subtype4	36	21	2.6 - 176.6 (44.2)

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

'RPPA CNMF subtypes' versus 'AGE'

P value = 0.305 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	166	55.3 (15.8)
subtype1	46	51.8 (13.4)
subtype2	58	55.3 (15.7)
subtype3	26	56.6 (17.1)
subtype4	36	58.8 (17.6)

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

'RPPA CNMF subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

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

Table S35. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	22	1	31	114
subtype1	2	0	14	31
subtype2	11	1	9	37
subtype3	4	0	5	17
subtype4	5	0	3	29

Figure S32. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'RPPA CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.559 (Fisher's exact test), Q value = 1

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

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	8	3	14	8	17	7	5	6	3	20	5	15	27	9
subtype1	2	1	5	2	7	1	2	1	0	3	3	5	9	2
subtype2	2	2	7	1	7	2	2	2	3	8	1	2	6	5
subtype3	1	0	1	2	1	0	0	1	0	5	1	4	5	0
subtype4	3	0	1	3	2	4	1	2	0	4	0	4	7	2

Figure S33. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.46 (Fisher's exact test), Q value = 1

Table S37. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	36	37	28	31
subtype1	8	12	10	9
subtype2	9	13	7	13
subtype3	7	7	5	2
subtype4	12	5	6	7

Figure S34. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.326 (Fisher's exact test), Q value = 1

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

nPatients	N0	N1	N2	N3
ALL	85	25	25	21
subtype1	22	7	6	8
subtype2	36	6	6	6
subtype3	8	6	7	3
subtype4	19	6	6	4

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

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

P value = 0.383 (Fisher's exact test), Q value = 1

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

nPatients	M0	M1	M1A	M1B	M1C
ALL	149	4	2	2	2
subtype1	42	1	1	0	0
subtype2	48	3	1	2	0
subtype3	25	0	0	0	0
subtype4	34	0	0	0	2

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

'RPPA CNMF subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.331 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	65	37
subtype1	21	7
subtype2	23	15
subtype3	11	5
subtype4	10	10

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

'RPPA CNMF subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

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

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

nPatients	NO	YES
ALL	25	144
subtype1	6	41
subtype2	8	50
subtype3	3	23
subtype4	8	30

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

'RPPA CNMF subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.224 (Kruskal-Wallis (anova)), Q value = 1

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

	nPatients	Mean (Std.Dev)
ALL	119	3.4 (5.3)
subtype1	36	3.6 (8.2)
subtype2	40	3.3 (3.0)
subtype3	20	2.6 (3.9)
subtype4	23	3.8 (3.5)

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

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.839 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	64	105
subtype1	17	30
subtype2	20	38
subtype3	11	15
subtype4	16	22

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

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S44. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	27	58	35	49

'RPPA cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.312 (logrank test), Q value = 1

Table S45. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	161	89	0.2 - 124.3 (14.8)
subtype1	25	14	1.4 - 84.7 (10.0)
subtype2	57	30	1.5 - 124.3 (23.9)
subtype3	34	20	1.1 - 82.3 (13.6)
subtype4	45	25	0.2 - 111.9 (11.7)

Figure S41. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.762 (logrank test), Q value = 1

Table S46. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	165	90	0.2 - 357.4 (50.9)
subtype1	25	14	9.9 - 215.4 (46.8)
subtype2	57	30	6.4 - 357.4 (53.9)
subtype3	34	20	11.3 - 346.0 (51.7)
subtype4	49	26	0.2 - 248.7 (48.9)

Figure S42. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

'RPPA cHierClus subtypes' versus 'AGE'

P value = 0.548 (Kruskal-Wallis (anova)), Q value = 1

Table S47. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	166	55.3 (15.8)
subtype1	26	58.5 (15.9)
subtype2	57	53.5 (15.2)
subtype3	34	56.4 (15.9)
subtype4	49	54.9 (16.6)

Figure S43. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'AGE'

'RPPA cHierClus subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.351 (Fisher's exact test), Q value = 1

Table S48. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	22	1	31	114
subtype1	3	0	8	16
subtype2	9	1	8	40
subtype3	2	0	4	28
subtype4	8	0	11	30

Figure S44. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'RPPA cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.3 (Fisher's exact test), Q value = 1

Table S49. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	8	3	14	8	17	7	5	6	3	20	5	15	27	9
subtype1	1	1	2	1	3	0	1	0	0	1	2	5	7	2
subtype2	2	1	7	3	5	0	3	4	3	10	0	2	9	2
subtype3	4	1	2	1	4	3	1	0	0	3	2	3	5	1
subtype4	1	0	3	3	5	4	0	2	0	6	1	5	6	4

Figure S45. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.302 (Fisher's exact test), Q value = 1

Table S50. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	36	37	28	31
subtype1	4	6	7	7
subtype2	16	12	4	11
subtype3	6	9	8	3
subtype4	10	10	9	10

Figure S46. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.579 (Fisher's exact test), Q value = 1

Table S51. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	85	25	25	21
subtype1	10	8	4	4
subtype2	32	5	8	9
subtype3	18	5	6	3
subtype4	25	7	7	5

Figure S47. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.811 (Fisher's exact test), Q value = 1

Table S52. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	149	4	2	2	2
subtype1	24	1	1	0	0
subtype2	51	1	0	1	0
subtype3	32	0	0	0	1
subtype4	42	2	1	1	1

Figure S48. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'RPPA cHierClus subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.843 (Fisher's exact test), Q value = 1

Table S53. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

nPatients	NO	YES
ALL	65	37
subtype1	11	6
subtype2	19	14
subtype3	12	5
subtype4	23	12

Figure S49. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

'RPPA cHierClus subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.132 (Fisher's exact test), Q value = 1

Table S54. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	25	144
subtype1	3	24
subtype2	14	44
subtype3	3	32
subtype4	5	44

Figure S50. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'RPPA cHierClus subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.882 (Kruskal-Wallis (anova)), Q value = 1

Table S55. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	119	3.4 (5.3)
subtype1	21	5.4 (10.7)
subtype2	36	3.1 (3.0)
subtype3	24	2.2 (1.8)
subtype4	38	3.2 (3.5)

Figure S51. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.275 (Fisher's exact test), Q value = 1

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	64	105
subtype1	6	21
subtype2	25	33
subtype3	15	20
subtype4	18	31

Figure S52. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

Clustering Approach #5: 'RNAseq CNMF subtypes'

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

Cluster Labels	1	2	3
Number of samples	97	83	105

'RNAseq CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 1.11e-05 (logrank test), Q value = 0.0015

Table S58. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	134	0.0 - 124.3 (14.1)
subtype1	93	52	1.6 - 122.7 (13.4)
subtype2	79	26	0.0 - 124.3 (27.4)
subtype3	100	56	0.0 - 114.7 (11.1)

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

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.000615 (logrank test), Q value = 0.078

Table S59. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	135	0.2 - 357.4 (47.5)
subtype1	93	52	4.9 - 357.4 (56.2)
subtype2	82	27	2.4 - 268.9 (53.9)
subtype3	104	56	0.2 - 247.0 (36.0)

Figure S54. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.00386 (Kruskal-Wallis (anova)), Q value = 0.47

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

	nPatients	Mean (Std.Dev)
ALL	280	55.7 (15.4)
subtype1	93	51.3 (15.2)
subtype2	83	57.1 (15.0)
subtype3	104	58.5 (15.2)

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

'RNAseq CNMF subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 5e-05 (Fisher's exact test), Q value = 0.0066

Table S61. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	39	4	61	180
subtype1	11	0	33	53
subtype2	5	0	13	65
subtype3	23	4	15	62

Figure S56. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'RNAseq CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.172 (Fisher's exact test), Q value = 1

Table S62. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	23	10	25	17	10	14	10	32	14	24	49	12
subtype1	4	3	10	1	9	7	4	4	2	12	5	6	14	7
subtype2	3	1	9	6	6	5	4	1	2	10	0	8	15	2
subtype3	3	2	4	3	10	5	2	9	6	10	9	10	20	3

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

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

P value = 0.0217 (Fisher's exact test), Q value = 1

Table S63. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	53	60	57	58
subtype1	18	29	17	13
subtype2	20	13	16	14
subtype3	15	18	24	31

Figure S58. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.799 (Fisher's exact test), Q value = 1

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

nPatients	N0	N1	N2	N3
ALL	141	52	34	37
subtype1	49	16	12	13
subtype2	43	16	6	10
subtype3	49	20	16	14

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

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

P value = 0.945 (Fisher's exact test), Q value = 1

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

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	2	5
subtype1	85	2	1	1	3
subtype2	73	1	0	0	1
subtype3	96	1	1	1	1

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

'RNAseq CNMF subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.214 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	104	74
subtype1	27	22
subtype2	35	16
subtype3	42	36

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

'RNAseq CNMF subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.0992 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	34	251
subtype1	15	82
subtype2	12	71
subtype3	7	98

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

'RNAseq CNMF subtypes' versus 'BRESLOW.THICKNESS'

P value = 8.04e-05 (Kruskal-Wallis (anova)), Q value = 0.011

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

	nPatients	Mean (Std.Dev)
ALL	208	3.6 (4.9)
subtype1	68	3.1 (6.3)
subtype2	56	2.8 (3.3)
subtype3	84	4.6 (4.5)

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

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.467 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	110	175
subtype1	42	55
subtype2	31	52
subtype3	37	68

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

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.528 (Fisher's exact test), Q value = 1

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

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	1	0	95
subtype2	1	1	80
subtype3	3	0	102

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

'RNAseq CNMF subtypes' versus 'ETHNICITY'

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

Table S71. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	276
subtype1	0	94
subtype2	0	82
subtype3	3	100

Figure S66. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S72. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	68	170	47

'RNAseq cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 4.84e-08 (logrank test), Q value = 6.6e-06

Table S73. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	134	0.0 - 124.3 (14.1)
subtype1	66	40	0.0 - 95.4 (8.6)
subtype2	161	66	0.0 - 124.3 (21.6)
subtype3	45	28	2.7 - 122.7 (13.0)

Figure S67. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 8.51e-05 (logrank test), Q value = 0.011

Table S74. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	135	0.2 - 357.4 (47.5)
subtype1	67	40	0.2 - 294.8 (28.7)
subtype2	167	67	0.2 - 346.0 (50.9)
subtype3	45	28	4.9 - 357.4 (53.9)

Figure S68. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.0021 (Kruskal-Wallis (anova)), Q value = 0.26

Table S75. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	280	55.7 (15.4)
subtype1	67	60.4 (15.5)
subtype2	168	55.5 (14.9)
subtype3	45	49.8 (15.2)

Figure S69. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'AGE'

'RNAseq cHierClus subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0013

Table S76. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	39	4	61	180
subtype1	22	3	14	28
subtype2	11	1	29	129
subtype3	6	0	18	23

Figure S70. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'RNAseq cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.173 (Fisher's exact test), Q value = 1

Table S77. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	23	10	25	17	10	14	10	32	14	24	49	12
subtype1	1	1	3	1	3	5	2	7	5	5	5	6	15	4
subtype2	6	2	17	9	16	8	6	7	4	22	5	15	26	6
subtype3	3	3	3	0	6	4	2	0	1	5	4	3	8	2

Figure S71. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.00036 (Fisher's exact test), Q value = 0.046

Table S78. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	53	60	57	58
subtype1	5	11	17	25
subtype2	42	34	30	28
subtype3	6	15	10	5

Figure S72. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.995 (Fisher's exact test), Q value = 1

Table S79. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	141	52	34	37
subtype1	32	12	8	10
subtype2	86	30	20	21
subtype3	23	10	6	6

Figure S73. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.223 (Fisher's exact test), Q value = 1

Table S80. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	2	5
subtype1	60	3	1	0	0
subtype2	152	1	1	1	4
subtype3	42	0	0	1	1

Figure S74. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'RNAseq cHierClus subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.12 (Fisher's exact test), Q value = 1

Table S81. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

nPatients	NO	YES
ALL	104	74
subtype1	23	26
subtype2	67	37
subtype3	14	11

Figure S75. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

'RNAseq cHierClus subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.342 (Fisher's exact test), Q value = 1

Table S82. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	34	251
subtype1	5	63
subtype2	24	146
subtype3	5	42

Figure S76. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'RNAseq cHierClus subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.000105 (Kruskal-Wallis (anova)), Q value = 0.014

Table S83. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	208	3.6 (4.9)
subtype1	55	4.9 (4.6)
subtype2	118	2.9 (3.3)
subtype3	35	3.9 (8.5)

Figure S77. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.633 (Fisher's exact test), Q value = 1

Table S84. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	110	175
subtype1	23	45
subtype2	69	101
subtype3	18	29

Figure S78. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.739 (Fisher's exact test), Q value = 1

Table S85. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	2	0	66
subtype2	2	1	165
subtype3	1	0	46

Figure S79. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.21 (Fisher's exact test), Q value = 1

Table S86. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	3	276
subtype1	2	65
subtype2	1	166
subtype3	0	45

Figure S80. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S87. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3	4	5
Number of samples	63	94	76	41	11

'MIRSEQ CNMF' versus 'Time from Specimen Diagnosis to Death'

P value = 0.277 (logrank test), Q value = 1

Table S88. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	136	0.0 - 142.4 (14.6)
subtype1	60	28	0.1 - 98.3 (15.3)
subtype2	89	51	0.0 - 142.4 (12.9)
subtype3	73	36	0.2 - 114.7 (13.2)
subtype4	39	16	0.2 - 122.7 (19.4)
subtype5	11	5	0.8 - 47.0 (25.1)

Figure S81. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.465 (logrank test), Q value = 1

Table S89. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	137	0.2 - 357.4 (47.5)
subtype1	61	28	6.3 - 346.0 (41.6)
subtype2	94	51	2.6 - 247.0 (48.0)
subtype3	73	36	0.2 - 314.6 (55.9)
subtype4	40	17	0.2 - 357.4 (40.9)
subtype5	11	5	10.0 - 194.8 (47.0)

Figure S82. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 0.121 (Kruskal-Wallis (anova)), Q value = 1

Table S90. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	280	55.6 (15.5)
subtype1	61	58.0 (14.1)
subtype2	94	56.9 (16.5)
subtype3	73	52.9 (14.2)
subtype4	41	56.3 (16.6)
subtype5	11	47.6 (15.1)

Figure S83. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'AGE'

'MIRSEQ CNMF' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.0089 (Fisher's exact test), Q value = 1

Table S91. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	41	4	60	179
subtype1	11	3	6	43
subtype2	13	1	14	65
subtype3	12	0	27	37
subtype4	4	0	9	28
subtype5	1	0	4	6

Figure S84. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'MIRSEQ CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0717 (Fisher's exact test), Q value = 1

Table S92. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	22	11	25	16	9	15	10	30	13	26	50	13
subtype1	5	5	8	6	3	3	1	1	3	6	1	2	12	0
subtype2	2	0	4	3	10	6	3	8	3	8	5	11	15	5
subtype3	2	1	8	1	11	2	3	4	2	9	5	6	14	3
subtype4	1	0	2	1	1	4	2	1	2	4	2	7	5	4
subtype5	0	0	0	0	0	1	0	1	0	3	0	0	4	1

Figure S85. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.00179 (Fisher's exact test), Q value = 0.22

Table S93. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	52	61	57	56
subtype1	16	9	6	12
subtype2	16	18	22	21
subtype3	6	29	16	12
subtype4	9	4	10	10
subtype5	5	1	3	1

Figure S86. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.08 (Fisher's exact test), Q value = 1

Table S94. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	142	51	34	38
subtype1	35	5	6	10
subtype2	48	16	12	12
subtype3	39	18	7	8
subtype4	17	8	9	4
subtype5	3	4	0	4

Figure S87. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.612 (Fisher's exact test), Q value = 1

Table S95. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	3	5
subtype1	58	0	0	0	0
subtype2	83	2	1	1	2
subtype3	69	1	0	1	1
subtype4	34	1	1	1	1
subtype5	10	0	0	0	1

Figure S88. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'MIRSEQ CNMF' versus 'MELANOMA.ULCERATION'

P value = 0.401 (Fisher's exact test), Q value = 1

Table S96. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

nPatients	NO	YES
ALL	104	74
subtype1	20	14
subtype2	42	27
subtype3	30	17
subtype4	11	13
subtype5	1	3

Figure S89. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

'MIRSEQ CNMF' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.0281 (Fisher's exact test), Q value = 1

Table S97. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	36	249
subtype1	7	56
subtype2	11	83
subtype3	6	70
subtype4	7	34
subtype5	5	6

Figure S90. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'MIRSEQ CNMF' versus 'BRESLOW.THICKNESS'

P value = 0.0548 (Kruskal-Wallis (anova)), Q value = 1

Table S98. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	207	3.6 (5.1)
subtype1	40	3.6 (5.0)
subtype2	73	3.8 (4.2)
subtype3	61	3.4 (6.7)
subtype4	28	3.7 (3.2)
subtype5	5	4.0 (2.3)

Figure S91. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.383 (Fisher's exact test), Q value = 1

Table S99. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	105	180
subtype1	26	37
subtype2	37	57
subtype3	21	55
subtype4	16	25
subtype5	5	6

Figure S92. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'GENDER'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.19 (Fisher's exact test), Q value = 1

Table S100. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	3	1	59
subtype2	0	0	93
subtype3	1	0	74
subtype4	1	0	40
subtype5	0	0	11

Figure S93. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.692 (Fisher's exact test), Q value = 1

Table S101. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	4	275
subtype1	0	62
subtype2	2	90
subtype3	2	71
subtype4	0	41
subtype5	0	11

Figure S94. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S102. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3	4
Number of samples	76	78	36	95

'MIRSEQ CHIERARCHICAL' versus 'Time from Specimen Diagnosis to Death'

P value = 0.316 (logrank test), Q value = 1

Table S103. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	136	0.0 - 142.4 (14.6)
subtype1	73	41	0.1 - 124.3 (13.2)
subtype2	73	37	0.1 - 114.7 (15.2)
subtype3	34	14	0.2 - 122.7 (19.9)
subtype4	92	44	0.0 - 142.4 (13.6)

Figure S95. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.603 (logrank test), Q value = 1

Table S104. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	137	0.2 - 357.4 (47.5)
subtype1	76	41	0.2 - 314.6 (43.9)
subtype2	74	37	4.9 - 346.0 (53.7)
subtype3	35	15	0.2 - 357.4 (34.8)
subtype4	94	44	2.6 - 294.8 (49.2)

Figure S96. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.773 (Kruskal-Wallis (anova)), Q value = 1

Table S105. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	280	55.6 (15.5)
subtype1	76	57.2 (14.7)
subtype2	74	54.8 (13.7)
subtype3	36	56.2 (16.7)
subtype4	94	54.8 (16.9)

Figure S97. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.425 (Fisher's exact test), Q value = 1

Table S106. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	41	4	60	179
subtype1	13	2	15	46
subtype2	8	0	20	50
subtype3	4	0	11	21
subtype4	16	2	14	62

Figure S98. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.346 (Fisher's exact test), Q value = 1

Table S107. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	22	11	25	16	9	15	10	30	13	26	50	13
subtype1	1	3	5	2	6	5	4	4	4	10	5	5	15	1
subtype2	3	3	9	2	7	3	1	4	2	10	3	3	17	2
subtype3	1	0	1	1	1	3	2	1	1	3	1	9	4	4
subtype4	5	0	7	6	11	5	2	6	3	7	4	9	14	6

Figure S99. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.215 (Fisher's exact test), Q value = 1

Table S108. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	52	61	57	56
subtype1	11	16	22	15
subtype2	14	19	9	14
subtype3	8	3	10	8
subtype4	19	23	16	19

Figure S100. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.396 (Fisher's exact test), Q value = 1

Table S109. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	142	51	34	38
subtype1	35	15	11	9
subtype2	39	13	7	13
subtype3	13	8	8	4
subtype4	55	15	8	12

Figure S101. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.103 (Fisher's exact test), Q value = 1

Table S110. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	3	5
subtype1	71	0	1	0	0
subtype2	70	1	0	1	0
subtype3	30	1	1	1	1
subtype4	83	2	0	1	4

Figure S102. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'MELANOMA.ULCERATION'

P value = 0.539 (Fisher's exact test), Q value = 1

Table S111. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

nPatients	NO	YES
ALL	104	74
subtype1	33	22
subtype2	22	17
subtype3	10	12
subtype4	39	23

Figure S103. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

'MIRSEQ CHIERARCHICAL' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.207 (Fisher's exact test), Q value = 1

Table S112. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	36	249
subtype1	5	71
subtype2	13	65
subtype3	6	30
subtype4	12	83

Figure S104. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'MIRSEQ CHIERARCHICAL' versus 'BRESLOW.THICKNESS'

P value = 0.202 (Kruskal-Wallis (anova)), Q value = 1

Table S113. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	207	3.6 (5.1)
subtype1	59	3.4 (3.0)
subtype2	50	4.3 (8.3)
subtype3	24	3.3 (2.9)
subtype4	74	3.4 (4.1)

Figure S105. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.49 (Fisher's exact test), Q value = 1

Table S114. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	105	180
subtype1	25	51
subtype2	27	51
subtype3	12	24
subtype4	41	54

Figure S106. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.358 (Fisher's exact test), Q value = 1

Table S115. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	1	1	74
subtype2	1	0	76
subtype3	2	0	34
subtype4	1	0	93

Figure S107. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.669 (Fisher's exact test), Q value = 1

Table S116. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	4	275
subtype1	0	75
subtype2	2	73
subtype3	0	35
subtype4	2	92

Figure S108. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S117. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	99	101	85

'MIRseq Mature CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.0392 (logrank test), Q value = 1

Table S118. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	136	0.0 - 142.4 (14.6)
subtype1	94	39	0.0 - 122.7 (16.3)
subtype2	96	55	0.0 - 142.4 (13.2)
subtype3	82	42	0.2 - 114.7 (12.8)

Figure S109. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

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

P value = 0.0706 (logrank test), Q value = 1

Table S119. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	137	0.2 - 357.4 (47.5)
subtype1	96	40	0.2 - 357.4 (47.3)
subtype2	101	55	2.6 - 248.7 (43.4)
subtype3	82	42	0.2 - 314.6 (55.8)

Figure S110. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.255 (Kruskal-Wallis (anova)), Q value = 1

Table S120. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	280	55.6 (15.5)
subtype1	97	55.8 (14.6)
subtype2	101	57.3 (16.7)
subtype3	82	53.5 (14.7)

Figure S111. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'AGE'

'MIRseq Mature CNMF subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.00187 (Fisher's exact test), Q value = 0.23

Table S121. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	41	4	60	179
subtype1	12	2	17	68
subtype2	18	2	12	68
subtype3	11	0	31	43

Figure S112. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

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

P value = 0.304 (Fisher's exact test), Q value = 1

Table S122. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	22	11	25	16	9	15	10	30	13	26	50	13
subtype1	6	3	8	7	3	5	1	2	4	12	2	11	18	4
subtype2	2	2	5	3	11	6	3	9	4	10	6	10	16	4
subtype3	2	1	9	1	11	5	5	4	2	8	5	5	16	5

Figure S113. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.00012 (Fisher's exact test), Q value = 0.015

Table S123. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	52	61	57	56
subtype1	28	12	12	21
subtype2	17	19	24	23
subtype3	7	30	21	12

Figure S114. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.798 (Fisher's exact test), Q value = 1

Table S124. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	142	51	34	38
subtype1	43	16	12	17
subtype2	53	19	13	11
subtype3	46	16	9	10

Figure S115. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.988 (Fisher's exact test), Q value = 1

Table S125. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	3	5
subtype1	86	1	1	1	1
subtype2	93	1	1	1	2
subtype3	75	2	0	1	2

Figure S116. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'MIRseq Mature CNMF subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.84 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	104	74
subtype1	31	25
subtype2	42	27
subtype3	31	22

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

'MIRseq Mature CNMF subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.241 (Fisher's exact test), Q value = 1

Table S127. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	36	249
subtype1	17	82
subtype2	11	90
subtype3	8	77

Figure S118. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'MIRseq Mature CNMF subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.0949 (Kruskal-Wallis (anova)), Q value = 1

Table S128. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	207	3.6 (5.1)
subtype1	61	3.7 (4.6)
subtype2	78	3.9 (4.2)
subtype3	68	3.3 (6.3)

Figure S119. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.305 (Fisher's exact test), Q value = 1

Table S129. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	105	180
subtype1	41	58
subtype2	38	63
subtype3	26	59

Figure S120. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.0575 (Fisher's exact test), Q value = 1

Table S130. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	4	1	94
subtype2	0	0	100
subtype3	1	0	83

Figure S121. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.46 (Fisher's exact test), Q value = 1

Table S131. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	4	275
subtype1	0	98
subtype2	2	98
subtype3	2	79

Figure S122. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'ETHNICITY'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S132. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	119	86	55	25

'MIRseq Mature cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.0835 (logrank test), Q value = 1

Table S133. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	272	136	0.0 - 142.4 (14.6)
subtype1	114	57	0.0 - 142.4 (13.9)
subtype2	81	37	0.0 - 122.7 (16.0)
subtype3	53	32	0.2 - 110.8 (12.6)
subtype4	24	10	0.2 - 114.2 (21.5)

Figure S123. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time from Specimen Diagnosis to Death'

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

P value = 0.465 (logrank test), Q value = 1

Table S134. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	279	137	0.2 - 357.4 (47.5)
subtype1	119	57	2.6 - 248.7 (41.6)
subtype2	82	37	4.9 - 346.0 (53.7)
subtype3	53	32	0.2 - 314.6 (53.3)
subtype4	25	11	0.2 - 357.4 (46.4)

Figure S124. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.819 (Kruskal-Wallis (anova)), Q value = 1

Table S135. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	280	55.6 (15.5)
subtype1	119	56.4 (15.7)
subtype2	82	54.2 (14.3)
subtype3	54	56.1 (16.2)
subtype4	25	55.8 (16.9)

Figure S125. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'AGE'

'MIRseq Mature cHierClus subtypes' versus 'PRIMARY.SITE.OF.DISEASE'

P value = 0.00419 (Fisher's exact test), Q value = 0.51

Table S136. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

nPatients	DISTANT METASTASIS	PRIMARY TUMOR	REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE	REGIONAL LYMPH NODE
ALL	41	4	60	179
subtype1	17	4	13	85
subtype2	9	0	24	53
subtype3	12	0	17	25
subtype4	3	0	6	16

Figure S126. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PRIMARY.SITE.OF.DISEASE'

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

P value = 0.494 (Fisher's exact test), Q value = 1

Table S137. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

nPatients	I OR II NOS	STAGE 0	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE IIC	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV
ALL	10	6	22	11	25	16	9	15	10	30	13	26	50	13
subtype1	5	3	5	7	10	9	3	9	4	13	5	10	22	3
subtype2	3	3	9	2	7	2	1	4	2	13	3	7	17	3
subtype3	1	0	7	1	7	2	3	1	3	3	4	5	9	5
subtype4	1	0	1	1	1	3	2	1	1	1	1	4	2	2

Figure S127. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'NEOPLASM.DISEASESTAGE'

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

P value = 0.134 (Fisher's exact test), Q value = 1

Table S138. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	52	61	57	56
subtype1	23	21	28	25
subtype2	19	20	11	14
subtype3	5	18	12	12
subtype4	5	2	6	5

Figure S128. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY.T.STAGE'

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

P value = 0.932 (Fisher's exact test), Q value = 1

Table S139. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

nPatients	N0	N1	N2	N3
ALL	142	51	34	38
subtype1	63	20	14	16
subtype2	37	18	10	14
subtype3	29	10	6	6
subtype4	13	3	4	2

Figure S129. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'PATHOLOGY.N.STAGE'

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

P value = 0.278 (Fisher's exact test), Q value = 1

Table S140. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	M1A	M1B	M1C
ALL	254	4	2	3	5
subtype1	108	1	1	1	1
subtype2	77	0	0	1	2
subtype3	48	2	1	0	2
subtype4	21	1	0	1	0

Figure S130. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'PATHOLOGY.M.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'MELANOMA.ULCERATION'

P value = 0.902 (Fisher's exact test), Q value = 1

Table S141. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

nPatients	NO	YES
ALL	104	74
subtype1	48	32
subtype2	23	19
subtype3	25	16
subtype4	8	7

Figure S131. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'MELANOMA.ULCERATION'

'MIRseq Mature cHierClus subtypes' versus 'MELANOMA.PRIMARY.KNOWN'

P value = 0.104 (Fisher's exact test), Q value = 1

Table S142. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

nPatients	NO	YES
ALL	36	249
subtype1	13	106
subtype2	16	70
subtype3	3	52
subtype4	4	21

Figure S132. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'MELANOMA.PRIMARY.KNOWN'

'MIRseq Mature cHierClus subtypes' versus 'BRESLOW.THICKNESS'

P value = 0.176 (Kruskal-Wallis (anova)), Q value = 1

Table S143. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

	nPatients	Mean (Std.Dev)
ALL	207	3.6 (5.1)
subtype1	90	3.8 (4.3)
subtype2	53	3.9 (7.7)
subtype3	48	3.1 (3.0)
subtype4	16	2.7 (1.9)

Figure S133. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'BRESLOW.THICKNESS'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.753 (Fisher's exact test), Q value = 1

Table S144. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

nPatients	FEMALE	MALE
ALL	105	180
subtype1	47	72
subtype2	28	58
subtype3	20	35
subtype4	10	15

Figure S134. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.663 (Fisher's exact test), Q value = 1

Table S145. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	5	1	277
subtype1	2	1	115
subtype2	2	0	83
subtype3	0	0	55
subtype4	1	0	24

Figure S135. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.877 (Fisher's exact test), Q value = 1

Table S146. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	4	275
subtype1	2	115
subtype2	2	81
subtype3	0	54
subtype4	0	25

Figure S136. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'ETHNICITY'

Methods & Data

Input

Cluster data file = SKCM-TM.mergedcluster.txt
Clinical data file = SKCM-TM.merged_data.txt
Number of patients = 295
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