Correlation between aggregated molecular cancer subtypes and selected clinical features
Skin Cutaneous Melanoma (Metastatic)
28 January 2016  |  analyses__2016_01_28
Maintainer Information
Citation Information
doi:10.7908/C1M32V6G
Maintained by TCGA GDAC Team (Broad Institute/MD Anderson Cancer Center/Harvard Medical School)
Cite as Broad Institute TCGA Genome Data Analysis Center (2016): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1M32V6G
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 368 patients, 32 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 'Time to Death',  'PATHOLOGY_N_STAGE', and 'GENDER'.

  • 4 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time from Specimen Diagnosis to Death',  'Time to Death',  'YEARS_TO_BIRTH', and 'BRESLOW_THICKNESS'.

  • CNMF clustering analysis on RPPA data identified 5 subtypes that correlate to 'Time from Specimen Diagnosis to Death',  'Time to Death', and 'PATHOLOGY_N_STAGE'.

  • Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that correlate to 'Time from Specimen Diagnosis to Death',  'Time to Death',  'PATHOLOGY_T_STAGE', and 'RADIATION_THERAPY'.

  • CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that correlate to 'Time from Specimen Diagnosis to Death',  'Time to Death',  'YEARS_TO_BIRTH',  'PATHOLOGIC_STAGE', 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',  'YEARS_TO_BIRTH',  'PATHOLOGIC_STAGE',  'PATHOLOGY_T_STAGE', and 'BRESLOW_THICKNESS'.

  • 4 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 correlate to 'Time from Specimen Diagnosis to Death' and 'BRESLOW_THICKNESS'.

  • 3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'PATHOLOGIC_STAGE',  'PATHOLOGY_T_STAGE', and 'BRESLOW_THICKNESS'.

  • 9 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'PATHOLOGIC_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, 32 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 0.393
(0.704) 		0.0034
(0.034) 		0.00192
(0.0224) 		0.00828
(0.0682) 		0.000103
(0.00287) 		1.99e-08
(1.58e-06) 		0.307
(0.652) 		0.0406
(0.184) 		0.959
(0.988) 		0.159
(0.429)
Time to Death logrank test 0.0337
(0.163) 		0.0171
(0.12) 		4.29e-06
(0.000152) 		0.042
(0.184) 		0.00024
(0.00479) 		2.25e-08
(1.58e-06) 		0.895
(0.957) 		0.0832
(0.303) 		0.661
(0.87) 		0.111
(0.354)
YEARS TO BIRTH Kruskal-Wallis (anova) 0.13
(0.397) 		0.0024
(0.0259) 		0.785
(0.909) 		0.081
(0.303) 		0.0207
(0.123) 		0.00151
(0.0212) 		0.319
(0.652) 		0.879
(0.955) 		0.0991
(0.347) 		0.881
(0.955)
PATHOLOGIC STAGE Fisher's exact test 0.73
(0.891) 		0.312
(0.652) 		0.755
(0.891) 		0.684
(0.885) 		0.0184
(0.123) 		0.0412
(0.184) 		0.482
(0.749) 		0.262
(0.582) 		0.0018
(0.0224) 		0.0203
(0.123)
PATHOLOGY T STAGE Fisher's exact test 0.397
(0.704) 		0.158
(0.429) 		0.707
(0.891) 		0.00793
(0.0682) 		0.0567
(0.234) 		0.00071
(0.011) 		0.0164
(0.12) 		0.192
(0.472) 		0.00068
(0.011) 		0.28
(0.613)
PATHOLOGY N STAGE Fisher's exact test 0.0239
(0.129) 		0.44
(0.734) 		0.0213
(0.123) 		0.0501
(0.213) 		0.844
(0.945) 		0.748
(0.891) 		0.422
(0.72) 		0.177
(0.467) 		0.188
(0.47) 		0.649
(0.87)
PATHOLOGY M STAGE Fisher's exact test 0.523
(0.779) 		0.516
(0.777) 		0.689
(0.885) 		0.475
(0.748) 		0.107
(0.35) 		0.96
(0.988) 		0.396
(0.704) 		0.33
(0.652) 		0.722
(0.891) 		0.18
(0.467)
MELANOMA ULCERATION Fisher's exact test 0.158
(0.429) 		0.242
(0.547) 		0.887
(0.955) 		0.918
(0.966) 		0.558
(0.814) 		0.543
(0.8) 		0.147
(0.421) 		0.446
(0.734) 		0.139
(0.406) 		0.645
(0.87)
MELANOMA PRIMARY KNOWN Fisher's exact test 0.508
(0.773) 		0.0662
(0.265) 		0.461
(0.742) 		0.185
(0.47) 		0.106
(0.35) 		0.417
(0.72) 		0.438
(0.734) 		0.468
(0.744) 		0.222
(0.522) 		0.971
(0.993)
BRESLOW THICKNESS Kruskal-Wallis (anova) 0.389
(0.704) 		0.0219
(0.123) 		0.405
(0.709) 		0.451
(0.734) 		0.000237
(0.00479) 		4.36e-06
(0.000152) 		0.105
(0.35) 		0.0335
(0.163) 		0.00437
(0.0408) 		0.133
(0.397)
GENDER Fisher's exact test 0.0333
(0.163) 		0.2
(0.483) 		0.296
(0.639) 		0.33
(0.652) 		0.84
(0.945) 		0.658
(0.87) 		0.752
(0.891) 		0.361
(0.701) 		0.382
(0.704) 		0.571
(0.824)
RADIATION THERAPY Fisher's exact test 0.799
(0.916) 		0.632
(0.87) 		0.665
(0.87) 		0.0116
(0.0905) 		0.628
(0.87) 		0.389
(0.704) 		0.133
(0.397) 		0.605
(0.865) 		0.37
(0.704) 		0.0843
(0.303)
RACE Fisher's exact test 0.497
(0.765) 		0.711
(0.891) 		0.0779
(0.303) 		0.224
(0.522) 		0.861
(0.949) 		0.94
(0.982) 		0.326
(0.652) 		0.237
(0.543) 		0.764
(0.891) 		0.856
(0.949)
ETHNICITY Fisher's exact test 0.806
(0.918) 		0.742
(0.891) 		0.739
(0.891) 		1
(1.00) 		0.643
(0.87) 		0.627
(0.87) 		0.913
(0.966) 		1
(1.00) 		1
(1.00) 		0.76
(0.891)
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 154 110 103
'Copy Number Ratio CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.393 (logrank test), Q value = 0.7

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 252 124 0.0 - 346.5 (46.5)
subtype1 104 58 0.0 - 346.5 (57.2)
subtype2 71 37 0.2 - 297.0 (28.7)
subtype3 77 29 0.2 - 201.6 (44.0)

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 = 0.0337 (logrank test), Q value = 0.16

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

nPatients nDeath Duration Range (Median), Month
ALL 358 192 0.2 - 369.9 (52.3)
subtype1 151 92 0.3 - 369.9 (50.1)
subtype2 105 57 0.2 - 340.1 (48.6)
subtype3 102 43 0.2 - 228.6 (62.6)

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 'YEARS_TO_BIRTH'

P value = 0.13 (Kruskal-Wallis (anova)), Q value = 0.4

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

nPatients Mean (Std.Dev)
ALL 359 56.3 (15.7)
subtype1 152 57.8 (16.0)
subtype2 105 53.4 (16.8)
subtype3 102 57.1 (13.7)

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.73 (Fisher's exact test), Q value = 0.89

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

nPatients I/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 13 7 29 18 28 26 14 19 15 39 15 35 54 21
subtype1 5 1 8 10 13 9 4 8 8 15 7 14 24 11
subtype2 4 3 8 4 6 10 3 4 6 14 6 12 15 5
subtype3 4 3 13 4 9 7 7 7 1 10 2 9 15 5

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.397 (Fisher's exact test), Q value = 0.7

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

nPatients T0+T1 T2 T3 T4
ALL 64 73 81 67
subtype1 33 29 29 30
subtype2 15 21 25 23
subtype3 16 23 27 14

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0239 (Fisher's exact test), Q value = 0.13

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

nPatients N0 N1 N2 N3
ALL 177 65 39 45
subtype1 72 22 19 24
subtype2 46 30 11 8
subtype3 59 13 9 13

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.523 (Fisher's exact test), Q value = 0.78

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

nPatients 0 1
ALL 319 22
subtype1 131 12
subtype2 96 5
subtype3 92 5

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

'Copy Number Ratio CNMF subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.158 (Fisher's exact test), Q value = 0.43

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

nPatients NO YES
ALL 134 90
subtype1 61 42
subtype2 31 29
subtype3 42 19

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

'Copy Number Ratio CNMF subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.508 (Fisher's exact test), Q value = 0.77

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

nPatients NO YES
ALL 47 320
subtype1 23 131
subtype2 11 99
subtype3 13 90

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

'Copy Number Ratio CNMF subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.389 (Kruskal-Wallis (anova)), Q value = 0.7

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

nPatients Mean (Std.Dev)
ALL 266 3.5 (4.7)
subtype1 109 3.3 (3.4)
subtype2 84 4.1 (6.7)
subtype3 73 3.0 (3.4)

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

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.0333 (Fisher's exact test), Q value = 0.16

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

nPatients FEMALE MALE
ALL 138 229
subtype1 63 91
subtype2 47 63
subtype3 28 75

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

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.799 (Fisher's exact test), Q value = 0.92

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

nPatients NO YES
ALL 319 47
subtype1 136 18
subtype2 94 16
subtype3 89 13

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

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.497 (Fisher's exact test), Q value = 0.76

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 353
subtype1 1 0 149
subtype2 2 0 105
subtype3 2 1 99

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 = 0.806 (Fisher's exact test), Q value = 0.92

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 352
subtype1 4 147
subtype2 1 105
subtype3 2 100

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 111 122 35 99
'METHLYATION CNMF' versus 'Time from Specimen Diagnosis to Death'

P value = 0.0034 (logrank test), Q value = 0.034

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

nPatients nDeath Duration Range (Median), Month
ALL 252 125 0.0 - 346.5 (46.5)
subtype1 78 46 0.0 - 174.1 (42.9)
subtype2 88 35 1.9 - 346.5 (40.6)
subtype3 26 17 0.2 - 196.0 (34.6)
subtype4 60 27 0.8 - 312.1 (62.2)

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 = 0.0171 (logrank test), Q value = 0.12

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

nPatients nDeath Duration Range (Median), Month
ALL 358 193 0.2 - 369.9 (52.3)
subtype1 106 67 0.3 - 228.6 (54.5)
subtype2 120 51 0.2 - 369.9 (49.5)
subtype3 35 24 0.2 - 247.0 (48.0)
subtype4 97 51 0.7 - 368.8 (54.5)

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

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.0024 (Kruskal-Wallis (anova)), Q value = 0.026

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

nPatients Mean (Std.Dev)
ALL 359 56.3 (15.7)
subtype1 106 61.0 (15.0)
subtype2 121 55.1 (14.9)
subtype3 35 54.5 (16.7)
subtype4 97 53.3 (16.2)

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

'METHLYATION CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.312 (Fisher's exact test), Q value = 0.65

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

nPatients I/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 13 7 29 18 28 25 14 19 15 39 15 35 55 21
subtype1 3 3 5 7 9 6 6 8 7 13 8 5 14 7
subtype2 4 2 12 5 8 8 3 3 1 15 5 13 24 7
subtype3 1 0 3 1 2 3 2 4 4 1 1 4 6 0
subtype4 5 2 9 5 9 8 3 4 3 10 1 13 11 7

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

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.158 (Fisher's exact test), Q value = 0.43

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

nPatients T0+T1 T2 T3 T4
ALL 64 73 80 68
subtype1 16 20 26 26
subtype2 27 23 28 18
subtype3 2 9 7 11
subtype4 19 21 19 13

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

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.44 (Fisher's exact test), Q value = 0.73

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

nPatients N0 N1 N2 N3
ALL 177 66 39 45
subtype1 54 20 12 12
subtype2 51 24 11 22
subtype3 18 6 2 3
subtype4 54 16 14 8

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

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.516 (Fisher's exact test), Q value = 0.78

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

nPatients 0 1
ALL 319 22
subtype1 96 8
subtype2 105 7
subtype3 31 0
subtype4 87 7

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

'METHLYATION CNMF' versus 'MELANOMA_ULCERATION'

P value = 0.242 (Fisher's exact test), Q value = 0.55

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

nPatients NO YES
ALL 134 91
subtype1 51 33
subtype2 47 23
subtype3 11 12
subtype4 25 23

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

'METHLYATION CNMF' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.0662 (Fisher's exact test), Q value = 0.26

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

nPatients NO YES
ALL 47 320
subtype1 10 101
subtype2 20 102
subtype3 1 34
subtype4 16 83

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

'METHLYATION CNMF' versus 'BRESLOW_THICKNESS'

P value = 0.0219 (Kruskal-Wallis (anova)), Q value = 0.12

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

nPatients Mean (Std.Dev)
ALL 267 3.5 (4.7)
subtype1 82 3.6 (3.2)
subtype2 85 3.7 (6.5)
subtype3 29 4.1 (3.9)
subtype4 71 2.8 (3.9)

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

'METHLYATION CNMF' versus 'GENDER'

P value = 0.2 (Fisher's exact test), Q value = 0.48

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

nPatients FEMALE MALE
ALL 137 230
subtype1 34 77
subtype2 49 73
subtype3 11 24
subtype4 43 56

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

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.632 (Fisher's exact test), Q value = 0.87

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

nPatients NO YES
ALL 318 48
subtype1 98 13
subtype2 102 20
subtype3 31 3
subtype4 87 12

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

'METHLYATION CNMF' versus 'RACE'

P value = 0.711 (Fisher's exact test), Q value = 0.89

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 353
subtype1 2 0 107
subtype2 1 1 117
subtype3 1 0 33
subtype4 1 0 96

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

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.742 (Fisher's exact test), Q value = 0.89

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 352
subtype1 1 106
subtype2 3 117
subtype3 1 34
subtype4 2 95

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 39 78 43 69 33
'RPPA CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.00192 (logrank test), Q value = 0.022

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 181 89 0.0 - 346.5 (44.3)
subtype1 33 5 3.9 - 346.5 (60.5)
subtype2 51 27 0.0 - 227.0 (31.9)
subtype3 27 16 0.2 - 150.6 (43.5)
subtype4 48 25 0.8 - 312.1 (49.3)
subtype5 22 16 4.4 - 297.0 (38.2)

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 = 4.29e-06 (logrank test), Q value = 0.00015

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

nPatients nDeath Duration Range (Median), Month
ALL 255 133 0.2 - 369.9 (50.1)
subtype1 39 6 7.0 - 369.9 (73.9)
subtype2 75 41 0.2 - 248.7 (41.6)
subtype3 41 28 2.6 - 176.6 (44.9)
subtype4 68 36 2.0 - 368.8 (57.2)
subtype5 32 22 3.6 - 297.9 (37.2)

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

'RPPA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.785 (Kruskal-Wallis (anova)), Q value = 0.91

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

nPatients Mean (Std.Dev)
ALL 256 56.1 (15.9)
subtype1 39 54.4 (16.4)
subtype2 76 56.6 (15.4)
subtype3 41 58.7 (15.8)
subtype4 68 54.9 (16.5)
subtype5 32 56.5 (15.3)

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

'RPPA CNMF subtypes' versus 'PATHOLOGIC_STAGE'

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

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

nPatients I/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 11 4 18 13 19 10 10 11 9 28 8 27 45 18
subtype1 4 0 3 3 3 2 1 0 0 5 1 4 6 2
subtype2 1 1 7 3 8 1 3 3 2 6 5 9 19 5
subtype3 3 0 1 3 3 4 1 3 2 6 1 3 5 2
subtype4 2 2 6 2 5 2 3 5 3 7 1 5 9 7
subtype5 1 1 1 2 0 1 2 0 2 4 0 6 6 2

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.707 (Fisher's exact test), Q value = 0.89

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

nPatients T0+T1 T2 T3 T4
ALL 50 46 52 46
subtype1 10 7 7 4
subtype2 12 18 19 15
subtype3 13 5 8 8
subtype4 9 11 12 14
subtype5 6 5 6 5

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0213 (Fisher's exact test), Q value = 0.12

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

nPatients N0 N1 N2 N3
ALL 113 43 33 38
subtype1 18 5 4 5
subtype2 30 21 7 13
subtype3 21 4 9 4
subtype4 37 7 5 12
subtype5 7 6 8 4

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

'RPPA CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.689 (Fisher's exact test), Q value = 0.88

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

nPatients 0 1
ALL 220 19
subtype1 32 2
subtype2 66 5
subtype3 37 2
subtype4 55 8
subtype5 30 2

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

'RPPA CNMF subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.887 (Fisher's exact test), Q value = 0.95

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

nPatients NO YES
ALL 91 66
subtype1 14 7
subtype2 28 22
subtype3 13 11
subtype4 26 17
subtype5 10 9

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

'RPPA CNMF subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.461 (Fisher's exact test), Q value = 0.74

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

nPatients NO YES
ALL 39 223
subtype1 8 31
subtype2 11 67
subtype3 8 35
subtype4 10 59
subtype5 2 31

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

'RPPA CNMF subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.405 (Kruskal-Wallis (anova)), Q value = 0.71

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

nPatients Mean (Std.Dev)
ALL 178 3.5 (4.8)
subtype1 23 2.8 (4.1)
subtype2 58 3.8 (6.8)
subtype3 27 3.6 (3.1)
subtype4 47 3.5 (3.0)
subtype5 23 3.5 (4.4)

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

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.296 (Fisher's exact test), Q value = 0.64

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

nPatients FEMALE MALE
ALL 106 156
subtype1 22 17
subtype2 29 49
subtype3 17 26
subtype4 25 44
subtype5 13 20

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

'RPPA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.665 (Fisher's exact test), Q value = 0.87

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

nPatients NO YES
ALL 228 34
subtype1 31 8
subtype2 69 9
subtype3 38 5
subtype4 60 9
subtype5 30 3

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

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.0779 (Fisher's exact test), Q value = 0.3

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

nPatients ASIAN WHITE
ALL 4 251
subtype1 1 38
subtype2 0 75
subtype3 0 43
subtype4 1 66
subtype5 2 29

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

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.739 (Fisher's exact test), Q value = 0.89

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 5 252
subtype1 0 39
subtype2 2 74
subtype3 0 42
subtype4 2 66
subtype5 1 31

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
Number of samples 86 84 92
'RPPA cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.00828 (logrank test), Q value = 0.068

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

nPatients nDeath Duration Range (Median), Month
ALL 181 89 0.0 - 346.5 (44.3)
subtype1 64 26 4.5 - 346.5 (56.5)
subtype2 56 29 0.0 - 225.1 (26.8)
subtype3 61 34 0.8 - 297.0 (47.2)

Figure S43.  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 = 0.042 (logrank test), Q value = 0.18

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

nPatients nDeath Duration Range (Median), Month
ALL 255 133 0.2 - 369.9 (50.1)
subtype1 84 40 6.4 - 369.9 (58.0)
subtype2 80 43 0.2 - 226.0 (36.0)
subtype3 91 50 2.0 - 368.8 (49.5)

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

'RPPA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.081 (Kruskal-Wallis (anova)), Q value = 0.3

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

nPatients Mean (Std.Dev)
ALL 256 56.1 (15.9)
subtype1 84 53.0 (15.9)
subtype2 81 58.7 (15.4)
subtype3 91 56.7 (15.9)

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

'RPPA cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.684 (Fisher's exact test), Q value = 0.88

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

nPatients I/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 11 4 18 13 19 10 10 11 9 28 8 27 45 18
subtype1 5 2 10 3 7 2 3 3 5 10 2 5 15 4
subtype2 4 0 4 4 6 5 3 4 0 7 4 12 17 6
subtype3 2 2 4 6 6 3 4 4 4 11 2 10 13 8

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

'RPPA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00793 (Fisher's exact test), Q value = 0.068

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

nPatients T0+T1 T2 T3 T4
ALL 50 46 52 46
subtype1 22 16 7 17
subtype2 11 15 28 13
subtype3 17 15 17 16

Figure S47.  Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0501 (Fisher's exact test), Q value = 0.21

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

nPatients N0 N1 N2 N3
ALL 113 43 33 38
subtype1 43 7 8 17
subtype2 31 21 12 10
subtype3 39 15 13 11

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

'RPPA cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.475 (Fisher's exact test), Q value = 0.75

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

nPatients 0 1
ALL 220 19
subtype1 72 4
subtype2 71 6
subtype3 77 9

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

'RPPA cHierClus subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.918 (Fisher's exact test), Q value = 0.97

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

nPatients NO YES
ALL 91 66
subtype1 27 21
subtype2 28 21
subtype3 36 24

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

'RPPA cHierClus subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.185 (Fisher's exact test), Q value = 0.47

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

nPatients NO YES
ALL 39 223
subtype1 18 68
subtype2 10 74
subtype3 11 81

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

'RPPA cHierClus subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.451 (Kruskal-Wallis (anova)), Q value = 0.73

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

nPatients Mean (Std.Dev)
ALL 178 3.5 (4.8)
subtype1 56 3.2 (3.7)
subtype2 57 3.9 (6.9)
subtype3 65 3.5 (3.4)

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

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.33 (Fisher's exact test), Q value = 0.65

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

nPatients FEMALE MALE
ALL 106 156
subtype1 40 46
subtype2 33 51
subtype3 33 59

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

'RPPA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.0116 (Fisher's exact test), Q value = 0.091

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

nPatients NO YES
ALL 228 34
subtype1 67 19
subtype2 76 8
subtype3 85 7

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

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.224 (Fisher's exact test), Q value = 0.52

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

nPatients ASIAN WHITE
ALL 4 251
subtype1 0 86
subtype2 1 80
subtype3 3 85

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 5 252
subtype1 2 84
subtype2 1 81
subtype3 2 87

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 4
Number of samples 117 73 105 73
'RNAseq CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.000103 (logrank test), Q value = 0.0029

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

nPatients nDeath Duration Range (Median), Month
ALL 253 125 0.0 - 346.5 (46.0)
subtype1 81 47 0.0 - 174.1 (30.9)
subtype2 56 18 0.3 - 201.6 (50.8)
subtype3 68 37 0.8 - 346.5 (60.6)
subtype4 48 23 0.8 - 297.0 (56.5)

Figure S57.  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 = 0.00024 (logrank test), Q value = 0.0048

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

nPatients nDeath Duration Range (Median), Month
ALL 359 193 0.2 - 369.9 (51.4)
subtype1 114 71 0.2 - 228.6 (46.7)
subtype2 72 28 0.3 - 222.5 (56.0)
subtype3 101 57 0.2 - 369.9 (62.1)
subtype4 72 37 0.7 - 302.1 (51.0)

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

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0207 (Kruskal-Wallis (anova)), Q value = 0.12

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

nPatients Mean (Std.Dev)
ALL 360 56.3 (15.7)
subtype1 114 59.7 (14.6)
subtype2 73 57.2 (15.0)
subtype3 101 53.8 (16.0)
subtype4 72 53.5 (16.7)

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

'RNAseq CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0184 (Fisher's exact test), Q value = 0.12

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

nPatients I/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 13 7 29 18 28 26 14 19 15 39 15 35 55 21
subtype1 3 3 2 7 9 8 3 8 8 11 7 10 23 3
subtype2 3 1 8 5 4 5 4 1 2 9 0 10 14 2
subtype3 6 3 11 2 7 9 4 3 2 13 8 8 12 8
subtype4 1 0 8 4 8 4 3 7 3 6 0 7 6 8

Figure S60.  Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGIC_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0567 (Fisher's exact test), Q value = 0.23

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

nPatients T0+T1 T2 T3 T4
ALL 64 73 81 68
subtype1 17 17 29 32
subtype2 18 11 17 13
subtype3 16 26 23 14
subtype4 13 19 12 9

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

'RNAseq CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

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

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

nPatients N0 N1 N2 N3
ALL 177 66 39 45
subtype1 49 22 14 17
subtype2 34 12 7 11
subtype3 55 21 12 10
subtype4 39 11 6 7

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

'RNAseq CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.107 (Fisher's exact test), Q value = 0.35

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

nPatients 0 1
ALL 320 22
subtype1 104 4
subtype2 65 2
subtype3 90 8
subtype4 61 8

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

'RNAseq CNMF subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.558 (Fisher's exact test), Q value = 0.81

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

nPatients NO YES
ALL 134 91
subtype1 46 39
subtype2 28 17
subtype3 30 20
subtype4 30 15

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

'RNAseq CNMF subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.106 (Fisher's exact test), Q value = 0.35

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

nPatients NO YES
ALL 47 321
subtype1 8 109
subtype2 11 62
subtype3 17 88
subtype4 11 62

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

'RNAseq CNMF subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.000237 (Kruskal-Wallis (anova)), Q value = 0.0048

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

nPatients Mean (Std.Dev)
ALL 267 3.5 (4.7)
subtype1 90 4.2 (3.5)
subtype2 50 3.2 (4.2)
subtype3 72 3.2 (6.2)
subtype4 55 2.9 (4.4)

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

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.84 (Fisher's exact test), Q value = 0.94

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

nPatients FEMALE MALE
ALL 138 230
subtype1 40 77
subtype2 29 44
subtype3 41 64
subtype4 28 45

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

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.628 (Fisher's exact test), Q value = 0.87

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

nPatients NO YES
ALL 319 48
subtype1 102 14
subtype2 60 13
subtype3 93 12
subtype4 64 9

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

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.861 (Fisher's exact test), Q value = 0.95

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 354
subtype1 2 0 113
subtype2 1 1 70
subtype3 1 0 102
subtype4 1 0 69

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

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.643 (Fisher's exact test), Q value = 0.87

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 353
subtype1 3 112
subtype2 0 72
subtype3 2 99
subtype4 2 70

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 106 165 97
'RNAseq cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 1.99e-08 (logrank test), Q value = 1.6e-06

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

nPatients nDeath Duration Range (Median), Month
ALL 253 125 0.0 - 346.5 (46.0)
subtype1 75 51 0.0 - 297.0 (25.6)
subtype2 116 42 0.3 - 269.0 (51.2)
subtype3 62 32 1.9 - 346.5 (62.2)

Figure S71.  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 = 2.25e-08 (logrank test), Q value = 1.6e-06

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

nPatients nDeath Duration Range (Median), Month
ALL 359 193 0.2 - 369.9 (51.4)
subtype1 103 74 0.2 - 297.9 (36.3)
subtype2 163 69 0.3 - 340.1 (55.7)
subtype3 93 50 0.2 - 369.9 (62.1)

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

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00151 (Kruskal-Wallis (anova)), Q value = 0.021

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

nPatients Mean (Std.Dev)
ALL 360 56.3 (15.7)
subtype1 103 59.3 (15.5)
subtype2 164 57.4 (15.2)
subtype3 93 51.1 (15.7)

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

'RNAseq cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0412 (Fisher's exact test), Q value = 0.18

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

nPatients I/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 13 7 29 18 28 26 14 19 15 39 15 35 55 21
subtype1 2 2 3 4 6 6 5 9 10 8 8 9 21 6
subtype2 5 2 16 10 15 10 6 8 3 19 2 19 24 10
subtype3 6 3 10 4 7 10 3 2 2 12 5 7 10 5

Figure S74.  Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGIC_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00071 (Fisher's exact test), Q value = 0.011

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

nPatients T0+T1 T2 T3 T4
ALL 64 73 81 68
subtype1 11 16 25 34
subtype2 34 34 33 26
subtype3 19 23 23 8

Figure S75.  Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.748 (Fisher's exact test), Q value = 0.89

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

nPatients N0 N1 N2 N3
ALL 177 66 39 45
subtype1 46 18 13 15
subtype2 82 27 18 21
subtype3 49 21 8 9

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

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.96 (Fisher's exact test), Q value = 0.99

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

nPatients 0 1
ALL 320 22
subtype1 92 6
subtype2 144 11
subtype3 84 5

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

'RNAseq cHierClus subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.543 (Fisher's exact test), Q value = 0.8

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

nPatients NO YES
ALL 134 91
subtype1 40 33
subtype2 65 42
subtype3 29 16

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

'RNAseq cHierClus subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.417 (Fisher's exact test), Q value = 0.72

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

nPatients NO YES
ALL 47 321
subtype1 10 96
subtype2 22 143
subtype3 15 82

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

'RNAseq cHierClus subtypes' versus 'BRESLOW_THICKNESS'

P value = 4.36e-06 (Kruskal-Wallis (anova)), Q value = 0.00015

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

nPatients Mean (Std.Dev)
ALL 267 3.5 (4.7)
subtype1 83 4.7 (4.4)
subtype2 118 3.0 (3.5)
subtype3 66 2.9 (6.3)

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

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.658 (Fisher's exact test), Q value = 0.87

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

nPatients FEMALE MALE
ALL 138 230
subtype1 39 67
subtype2 59 106
subtype3 40 57

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

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.389 (Fisher's exact test), Q value = 0.7

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

nPatients NO YES
ALL 319 48
subtype1 93 12
subtype2 139 26
subtype3 87 10

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

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.94 (Fisher's exact test), Q value = 0.98

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 354
subtype1 2 0 103
subtype2 2 1 157
subtype3 1 0 94

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

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.627 (Fisher's exact test), Q value = 0.87

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 353
subtype1 3 101
subtype2 2 160
subtype3 2 92

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 97 104 97 54
'MIRSEQ CNMF' versus 'Time from Specimen Diagnosis to Death'

P value = 0.307 (logrank test), Q value = 0.65

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

nPatients nDeath Duration Range (Median), Month
ALL 244 123 0.0 - 346.5 (46.5)
subtype1 69 34 0.2 - 346.5 (53.3)
subtype2 74 42 0.2 - 209.8 (45.6)
subtype3 69 35 0.3 - 269.0 (45.9)
subtype4 32 12 0.0 - 312.1 (41.0)

Figure S85.  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 = 0.895 (logrank test), Q value = 0.96

Table S93.  Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 343 187 0.2 - 369.9 (53.2)
subtype1 93 49 0.2 - 369.9 (47.5)
subtype2 103 60 0.7 - 268.7 (56.8)
subtype3 94 51 0.2 - 368.8 (60.0)
subtype4 53 27 7.1 - 357.4 (46.8)

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

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.319 (Kruskal-Wallis (anova)), Q value = 0.65

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

nPatients Mean (Std.Dev)
ALL 344 56.4 (15.7)
subtype1 94 57.5 (14.5)
subtype2 103 57.7 (16.6)
subtype3 94 54.2 (15.1)
subtype4 53 55.8 (16.8)

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

'MIRSEQ CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.482 (Fisher's exact test), Q value = 0.75

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

nPatients I/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 13 7 28 16 28 24 12 18 14 37 14 33 54 20
subtype1 6 5 9 10 3 7 2 2 5 10 2 6 15 3
subtype2 2 1 6 3 10 7 3 10 3 12 5 11 15 7
subtype3 3 1 9 2 11 5 4 5 4 11 4 9 16 5
subtype4 2 0 4 1 4 5 3 1 2 4 3 7 8 5

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

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.0164 (Fisher's exact test), Q value = 0.12

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

nPatients T0+T1 T2 T3 T4
ALL 60 73 76 63
subtype1 25 13 15 16
subtype2 18 20 27 20
subtype3 7 30 21 17
subtype4 10 10 13 10

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

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.422 (Fisher's exact test), Q value = 0.72

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

nPatients N0 N1 N2 N3
ALL 170 62 39 43
subtype1 48 11 6 16
subtype2 50 20 14 12
subtype3 47 20 10 10
subtype4 25 11 9 5

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

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.396 (Fisher's exact test), Q value = 0.7

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

nPatients 0 1
ALL 307 21
subtype1 84 3
subtype2 91 8
subtype3 86 5
subtype4 46 5

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

'MIRSEQ CNMF' versus 'MELANOMA_ULCERATION'

P value = 0.147 (Fisher's exact test), Q value = 0.42

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

nPatients NO YES
ALL 127 87
subtype1 33 19
subtype2 45 27
subtype3 36 22
subtype4 13 19

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

'MIRSEQ CNMF' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.438 (Fisher's exact test), Q value = 0.73

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

nPatients NO YES
ALL 45 307
subtype1 13 84
subtype2 13 91
subtype3 9 88
subtype4 10 44

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

'MIRSEQ CNMF' versus 'BRESLOW_THICKNESS'

P value = 0.105 (Kruskal-Wallis (anova)), Q value = 0.35

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

nPatients Mean (Std.Dev)
ALL 254 3.4 (4.7)
subtype1 59 3.3 (4.4)
subtype2 81 3.7 (4.1)
subtype3 75 3.4 (6.3)
subtype4 39 3.3 (2.9)

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

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.752 (Fisher's exact test), Q value = 0.89

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

nPatients FEMALE MALE
ALL 131 221
subtype1 39 58
subtype2 39 65
subtype3 32 65
subtype4 21 33

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

'MIRSEQ CNMF' versus 'RADIATION_THERAPY'

P value = 0.133 (Fisher's exact test), Q value = 0.4

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

nPatients NO YES
ALL 308 43
subtype1 84 12
subtype2 91 13
subtype3 90 7
subtype4 43 11

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

'MIRSEQ CNMF' versus 'RACE'

P value = 0.326 (Fisher's exact test), Q value = 0.65

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 338
subtype1 3 1 92
subtype2 0 0 101
subtype3 1 0 92
subtype4 1 0 53

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

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.913 (Fisher's exact test), Q value = 0.97

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 337
subtype1 1 95
subtype2 3 99
subtype3 2 90
subtype4 1 53

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 4
Number of samples 132 93 89 38
'MIRSEQ CHIERARCHICAL' versus 'Time from Specimen Diagnosis to Death'

P value = 0.0406 (logrank test), Q value = 0.18

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

nPatients nDeath Duration Range (Median), Month
ALL 244 123 0.0 - 346.5 (46.5)
subtype1 92 42 0.8 - 297.0 (53.9)
subtype2 63 33 1.9 - 346.5 (52.9)
subtype3 64 38 0.2 - 196.0 (33.9)
subtype4 25 10 0.0 - 312.1 (36.3)

Figure S99.  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 = 0.0832 (logrank test), Q value = 0.3

Table S108.  Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 343 187 0.2 - 369.9 (53.2)
subtype1 126 67 0.7 - 340.1 (58.7)
subtype2 91 46 0.2 - 369.9 (56.4)
subtype3 89 56 0.2 - 247.0 (47.5)
subtype4 37 18 7.1 - 357.4 (45.3)

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

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.879 (Kruskal-Wallis (anova)), Q value = 0.95

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

nPatients Mean (Std.Dev)
ALL 344 56.4 (15.7)
subtype1 126 56.7 (16.5)
subtype2 91 55.3 (14.9)
subtype3 89 56.9 (15.0)
subtype4 38 56.9 (16.5)

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

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGIC_STAGE'

P value = 0.262 (Fisher's exact test), Q value = 0.58

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

nPatients I/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 13 7 28 16 28 24 12 18 14 37 14 33 54 20
subtype1 8 0 11 9 14 6 5 8 5 11 4 9 22 10
subtype2 3 3 11 4 6 7 1 4 3 11 4 6 16 3
subtype3 1 4 4 2 7 7 3 5 5 12 5 9 13 3
subtype4 1 0 2 1 1 4 3 1 1 3 1 9 3 4

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

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.192 (Fisher's exact test), Q value = 0.47

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

nPatients T0+T1 T2 T3 T4
ALL 60 73 76 63
subtype1 21 36 23 25
subtype2 18 19 16 13
subtype3 13 14 26 18
subtype4 8 4 11 7

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

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 0.177 (Fisher's exact test), Q value = 0.47

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

nPatients N0 N1 N2 N3
ALL 170 62 39 43
subtype1 72 19 11 18
subtype2 47 19 6 12
subtype3 36 16 14 10
subtype4 15 8 8 3

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

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.33 (Fisher's exact test), Q value = 0.65

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

nPatients 0 1
ALL 307 21
subtype1 115 10
subtype2 82 3
subtype3 78 4
subtype4 32 4

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

'MIRSEQ CHIERARCHICAL' versus 'MELANOMA_ULCERATION'

P value = 0.446 (Fisher's exact test), Q value = 0.73

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

nPatients NO YES
ALL 127 87
subtype1 50 33
subtype2 29 18
subtype3 38 23
subtype4 10 13

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

'MIRSEQ CHIERARCHICAL' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.468 (Fisher's exact test), Q value = 0.74

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

nPatients NO YES
ALL 45 307
subtype1 16 116
subtype2 15 78
subtype3 8 81
subtype4 6 32

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

'MIRSEQ CHIERARCHICAL' versus 'BRESLOW_THICKNESS'

P value = 0.0335 (Kruskal-Wallis (anova)), Q value = 0.16

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

nPatients Mean (Std.Dev)
ALL 254 3.4 (4.7)
subtype1 102 3.3 (4.0)
subtype2 61 3.7 (7.4)
subtype3 65 3.5 (3.0)
subtype4 26 3.2 (2.8)

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

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.361 (Fisher's exact test), Q value = 0.7

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

nPatients FEMALE MALE
ALL 131 221
subtype1 57 75
subtype2 32 61
subtype3 30 59
subtype4 12 26

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

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY'

P value = 0.605 (Fisher's exact test), Q value = 0.86

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

nPatients NO YES
ALL 308 43
subtype1 118 14
subtype2 81 12
subtype3 78 10
subtype4 31 7

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

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.237 (Fisher's exact test), Q value = 0.54

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 338
subtype1 1 0 129
subtype2 1 0 90
subtype3 1 1 83
subtype4 2 0 36

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

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

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

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 337
subtype1 3 128
subtype2 2 87
subtype3 2 85
subtype4 0 37

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 85 119 134
'MIRseq Mature CNMF subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.959 (logrank test), Q value = 0.99

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

nPatients nDeath Duration Range (Median), Month
ALL 234 119 0.0 - 346.5 (47.6)
subtype1 59 25 4.5 - 185.1 (43.5)
subtype2 82 45 0.0 - 312.1 (53.3)
subtype3 93 49 0.2 - 346.5 (43.2)

Figure S113.  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 = 0.661 (logrank test), Q value = 0.87

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

nPatients nDeath Duration Range (Median), Month
ALL 329 179 0.2 - 369.9 (53.5)
subtype1 79 36 7.0 - 302.1 (49.5)
subtype2 117 67 0.2 - 368.8 (61.1)
subtype3 133 76 0.2 - 369.9 (49.5)

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

'MIRseq Mature CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0991 (Kruskal-Wallis (anova)), Q value = 0.35

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

nPatients Mean (Std.Dev)
ALL 330 56.5 (15.8)
subtype1 80 57.5 (12.9)
subtype2 117 54.0 (16.7)
subtype3 133 58.2 (16.4)

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

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0018 (Fisher's exact test), Q value = 0.022

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

nPatients I/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 13 7 28 16 27 24 12 17 14 37 13 31 49 18
subtype1 7 2 5 10 3 4 3 1 0 14 0 7 17 4
subtype2 2 2 12 2 14 8 6 5 4 8 7 11 18 8
subtype3 4 3 11 4 10 12 3 11 10 15 6 13 14 6

Figure S116.  Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGIC_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00068 (Fisher's exact test), Q value = 0.011

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

nPatients T0+T1 T2 T3 T4
ALL 58 72 73 58
subtype1 27 12 14 9
subtype2 12 33 29 18
subtype3 19 27 30 31

Figure S117.  Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.188 (Fisher's exact test), Q value = 0.47

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

nPatients N0 N1 N2 N3
ALL 168 61 35 38
subtype1 40 12 9 17
subtype2 59 22 14 11
subtype3 69 27 12 10

Figure S118.  Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.722 (Fisher's exact test), Q value = 0.89

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

nPatients 0 1
ALL 297 19
subtype1 76 5
subtype2 101 8
subtype3 120 6

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

'MIRseq Mature CNMF subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.139 (Fisher's exact test), Q value = 0.41

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

nPatients NO YES
ALL 122 82
subtype1 31 13
subtype2 44 27
subtype3 47 42

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

'MIRseq Mature CNMF subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.222 (Fisher's exact test), Q value = 0.52

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

nPatients NO YES
ALL 41 297
subtype1 15 70
subtype2 12 107
subtype3 14 120

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

'MIRseq Mature CNMF subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.00437 (Kruskal-Wallis (anova)), Q value = 0.041

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

nPatients Mean (Std.Dev)
ALL 245 3.4 (4.8)
subtype1 49 3.1 (4.7)
subtype2 94 3.3 (5.7)
subtype3 102 3.6 (3.7)

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

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.382 (Fisher's exact test), Q value = 0.7

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

nPatients FEMALE MALE
ALL 123 215
subtype1 35 50
subtype2 38 81
subtype3 50 84

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

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.37 (Fisher's exact test), Q value = 0.7

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

nPatients NO YES
ALL 298 39
subtype1 72 13
subtype2 105 14
subtype3 121 12

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

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.764 (Fisher's exact test), Q value = 0.89

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 325
subtype1 1 1 82
subtype2 2 0 113
subtype3 2 0 130

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

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 324
subtype1 2 82
subtype2 2 113
subtype3 3 129

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 4 5 6 7 8 9
Number of samples 36 24 51 64 38 17 64 29 15
'MIRseq Mature cHierClus subtypes' versus 'Time from Specimen Diagnosis to Death'

P value = 0.159 (logrank test), Q value = 0.43

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

nPatients nDeath Duration Range (Median), Month
ALL 234 119 0.0 - 346.5 (47.6)
subtype1 24 12 3.9 - 185.1 (60.1)
subtype2 16 11 4.5 - 146.6 (42.1)
subtype3 35 14 4.5 - 171.6 (32.5)
subtype4 47 28 0.2 - 297.0 (43.2)
subtype5 29 18 2.0 - 225.1 (64.8)
subtype6 9 4 8.1 - 346.5 (85.0)
subtype7 45 23 0.3 - 269.0 (46.0)
subtype8 19 6 0.0 - 312.1 (45.7)
subtype9 10 3 8.3 - 227.0 (72.5)

Figure S127.  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 = 0.111 (logrank test), Q value = 0.35

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

nPatients nDeath Duration Range (Median), Month
ALL 329 179 0.2 - 369.9 (53.5)
subtype1 35 17 7.0 - 219.2 (49.5)
subtype2 23 18 6.4 - 222.5 (50.9)
subtype3 51 23 3.6 - 176.6 (43.4)
subtype4 64 41 0.2 - 297.9 (53.6)
subtype5 37 21 0.2 - 226.0 (66.5)
subtype6 14 5 17.9 - 369.9 (102.9)
subtype7 61 35 0.3 - 340.1 (55.7)
subtype8 29 13 7.1 - 357.4 (46.8)
subtype9 15 6 17.0 - 248.7 (59.6)

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

'MIRseq Mature cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.881 (Kruskal-Wallis (anova)), Q value = 0.95

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

nPatients Mean (Std.Dev)
ALL 330 56.5 (15.8)
subtype1 35 59.2 (13.2)
subtype2 23 59.8 (15.8)
subtype3 51 57.9 (14.7)
subtype4 64 56.4 (14.6)
subtype5 37 54.8 (12.7)
subtype6 14 51.9 (17.5)
subtype7 62 55.5 (18.6)
subtype8 29 56.6 (17.4)
subtype9 15 53.6 (20.3)

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

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.0203 (Fisher's exact test), Q value = 0.12

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

nPatients I/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 13 7 28 16 27 24 12 17 14 37 13 31 49 18
subtype1 1 1 6 4 1 1 2 2 0 3 0 3 10 0
subtype2 1 0 0 1 1 2 0 2 1 2 4 2 3 3
subtype3 4 2 1 3 4 3 2 2 3 8 2 4 5 2
subtype4 0 2 5 2 7 6 0 7 3 8 2 8 7 1
subtype5 0 2 5 0 4 3 1 1 0 7 3 3 4 1
subtype6 4 0 0 1 0 2 0 0 1 2 0 2 0 1
subtype7 1 0 8 3 8 1 4 2 5 3 1 4 15 6
subtype8 1 0 1 1 1 4 3 0 1 2 1 4 3 3
subtype9 1 0 2 1 1 2 0 1 0 2 0 1 2 1

Figure S130.  Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGIC_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.28 (Fisher's exact test), Q value = 0.61

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

nPatients T0+T1 T2 T3 T4
ALL 58 72 73 58
subtype1 11 5 7 6
subtype2 3 4 8 4
subtype3 10 9 10 9
subtype4 11 12 15 14
subtype5 4 14 10 2
subtype6 4 1 2 1
subtype7 7 19 11 14
subtype8 6 3 7 5
subtype9 2 5 3 3

Figure S131.  Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_T_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.649 (Fisher's exact test), Q value = 0.87

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

nPatients N0 N1 N2 N3
ALL 168 61 35 38
subtype1 16 5 2 9
subtype2 11 5 4 2
subtype3 24 9 8 5
subtype4 34 11 6 7
subtype5 19 11 3 3
subtype6 9 4 0 1
subtype7 30 10 6 8
subtype8 13 4 6 2
subtype9 12 2 0 1

Figure S132.  Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_N_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.18 (Fisher's exact test), Q value = 0.47

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

nPatients 0 1
ALL 297 19
subtype1 34 0
subtype2 20 3
subtype3 44 3
subtype4 57 1
subtype5 35 1
subtype6 14 1
subtype7 55 6
subtype8 24 3
subtype9 14 1

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

'MIRseq Mature cHierClus subtypes' versus 'MELANOMA_ULCERATION'

P value = 0.645 (Fisher's exact test), Q value = 0.87

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

nPatients NO YES
ALL 122 82
subtype1 11 12
subtype2 12 4
subtype3 22 10
subtype4 26 15
subtype5 9 7
subtype6 3 1
subtype7 26 20
subtype8 8 9
subtype9 5 4

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

'MIRseq Mature cHierClus subtypes' versus 'MELANOMA_PRIMARY_KNOWN'

P value = 0.971 (Fisher's exact test), Q value = 0.99

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

nPatients NO YES
ALL 41 297
subtype1 5 31
subtype2 3 21
subtype3 5 46
subtype4 7 57
subtype5 4 34
subtype6 3 14
subtype7 7 57
subtype8 5 24
subtype9 2 13

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

'MIRseq Mature cHierClus subtypes' versus 'BRESLOW_THICKNESS'

P value = 0.133 (Kruskal-Wallis (anova)), Q value = 0.4

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

nPatients Mean (Std.Dev)
ALL 245 3.4 (4.8)
subtype1 27 4.2 (6.2)
subtype2 18 3.0 (2.0)
subtype3 33 3.3 (3.1)
subtype4 51 3.7 (4.5)
subtype5 30 3.0 (8.9)
subtype6 5 2.9 (3.0)
subtype7 50 3.2 (3.3)
subtype8 19 3.3 (3.1)
subtype9 12 2.6 (1.9)

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

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.571 (Fisher's exact test), Q value = 0.82

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

nPatients FEMALE MALE
ALL 123 215
subtype1 14 22
subtype2 7 17
subtype3 24 27
subtype4 20 44
subtype5 10 28
subtype6 8 9
subtype7 23 41
subtype8 12 17
subtype9 5 10

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

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.0843 (Fisher's exact test), Q value = 0.3

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

nPatients NO YES
ALL 298 39
subtype1 28 8
subtype2 23 1
subtype3 44 7
subtype4 56 7
subtype5 37 1
subtype6 15 2
subtype7 59 5
subtype8 22 7
subtype9 14 1

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

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.856 (Fisher's exact test), Q value = 0.95

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

nPatients ASIAN BLACK OR AFRICAN AMERICAN WHITE
ALL 5 1 325
subtype1 1 0 34
subtype2 0 0 23
subtype3 1 1 48
subtype4 1 0 61
subtype5 1 0 36
subtype6 0 0 17
subtype7 0 0 63
subtype8 1 0 28
subtype9 0 0 15

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 = 0.76 (Fisher's exact test), Q value = 0.89

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

nPatients HISPANIC OR LATINO NOT HISPANIC OR LATINO
ALL 7 324
subtype1 0 34
subtype2 0 24
subtype3 1 50
subtype4 3 60
subtype5 1 34
subtype6 1 16
subtype7 1 62
subtype8 0 29
subtype9 0 15

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/SKCM-TM/22546664/SKCM-TM.mergedcluster.txt

  • Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/SKCM-TM/22507034/SKCM-TM.merged_data.txt

  • Number of patients = 368

  • 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)
[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)
Copyright © 2016 Broad Institute TCGA GDAC as part of the TCGA Research Network. All rights reserved.
Made with Nozzle