Breast Invasive Carcinoma: Correlation between molecular cancer subtypes and selected clinical features
(primary solid tumor cohort)
Maintained by TCGA GDAC Team (Broad Institute/MD Anderson Cancer Center/Harvard Medical School)
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 9 clinical features across 873 patients, 11 significant findings detected with P value < 0.05 and Q value < 0.25.

  • CNMF clustering analysis on array-based mRNA expression data identified 8 subtypes that correlate to 'AGE'.

  • Consensus hierarchical clustering analysis on array-based mRNA expression data identified 3 subtypes that correlate to 'AGE'.

  • 5 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes correlate to 'GENDER'.

  • 6 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'AGE'.

  • CNMF clustering analysis on RPPA data identified 3 subtypes that do not correlate to any clinical features.

  • Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that correlate to 'AGE'.

  • CNMF clustering analysis on sequencing-based mRNA expression data identified 8 subtypes that correlate to 'Time to Death',  'AGE', and 'NUMBER.OF.LYMPH.NODES'.

  • Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'AGE' and 'NEOPLASM.DISEASESTAGE'.

  • 3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'NEOPLASM.DISEASESTAGE'.

  • 3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. 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 9 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 11 significant findings detected.

Clinical
Features 		Time
to
Death 		AGE GENDER RADIATIONS
RADIATION
REGIMENINDICATION 		DISTANT
METASTASIS 		LYMPH
NODE
METASTASIS 		NUMBER
OF
LYMPH
NODES 		TUMOR
STAGECODE 		NEOPLASM
DISEASESTAGE
Statistical Tests logrank test ANOVA Fisher's exact test Fisher's exact test Chi-square test Chi-square test ANOVA ANOVA Chi-square test
mRNA CNMF subtypes 0.159
(1.00) 		7.84e-07
(6.27e-05) 		0.0612
(1.00) 		0.921
(1.00) 		0.538
(1.00) 		0.06
(1.00) 		0.415
(1.00) 		0.036
(1.00)
mRNA cHierClus subtypes 0.609
(1.00) 		0.000437
(0.0323) 		0.493
(1.00) 		0.105
(1.00) 		0.173
(1.00) 		0.117
(1.00) 		0.692
(1.00) 		0.0779
(1.00)
Copy Number Ratio CNMF subtypes 0.376
(1.00) 		0.011
(0.738) 		0.000381
(0.0286) 		0.454
(1.00) 		0.689
(1.00) 		0.364
(1.00) 		0.123
(1.00) 		0.494
(1.00)
METHLYATION CNMF 0.162
(1.00) 		0.0015
(0.107) 		0.0712
(1.00) 		0.183
(1.00) 		0.745
(1.00) 		0.328
(1.00) 		0.273
(1.00) 		0.0398
(1.00)
RPPA CNMF subtypes 0.0477
(1.00) 		0.0287
(1.00) 		0.0564
(1.00) 		0.981
(1.00) 		0.323
(1.00) 		0.0979
(1.00) 		0.0305
(1.00) 		0.0292
(1.00)
RPPA cHierClus subtypes 0.318
(1.00) 		0.00121
(0.0868) 		0.154
(1.00) 		0.917
(1.00) 		0.665
(1.00) 		0.194
(1.00) 		0.0553
(1.00) 		0.0496
(1.00)
RNAseq CNMF subtypes 0.000437
(0.0323) 		2.07e-05
(0.00161) 		0.0993
(1.00) 		0.475
(1.00) 		0.74
(1.00) 		0.0188
(1.00) 		0.00172
(0.12) 		0.00611
(0.416)
RNAseq cHierClus subtypes 0.0339
(1.00) 		3.71e-05
(0.00285) 		0.0115
(0.761) 		0.239
(1.00) 		0.3
(1.00) 		0.041
(1.00) 		0.919
(1.00) 		1.43e-05
(0.00113)
MIRSEQ CNMF 0.0226
(1.00) 		0.0567
(1.00) 		0.327
(1.00) 		0.243
(1.00) 		0.697
(1.00) 		0.00385
(0.266) 		0.141
(1.00) 		4.74e-05
(0.0036)
MIRSEQ CHIERARCHICAL 0.33
(1.00) 		0.0651
(1.00) 		0.187
(1.00) 		0.285
(1.00) 		0.554
(1.00) 		0.51
(1.00) 		0.804
(1.00) 		0.593
(1.00)
Clustering Approach #1: 'mRNA CNMF subtypes'

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

Cluster Labels 1 2 3 4 5 6 7 8
Number of samples 126 39 23 100 74 20 104 40
'mRNA CNMF subtypes' versus 'Time to Death'

P value = 0.159 (logrank test), Q value = 1

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

nPatients nDeath Duration Range (Median), Month
ALL 494 65 0.1 - 223.4 (24.2)
subtype1 116 17 0.1 - 177.4 (24.4)
subtype2 38 3 0.1 - 157.4 (43.5)
subtype3 21 3 0.3 - 223.4 (14.0)
subtype4 95 12 0.1 - 211.5 (22.6)
subtype5 65 14 0.1 - 189.0 (24.6)
subtype6 19 2 0.2 - 97.5 (36.3)
subtype7 101 10 0.3 - 220.9 (19.0)
subtype8 39 4 0.3 - 112.4 (23.4)

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

'mRNA CNMF subtypes' versus 'AGE'

P value = 7.84e-07 (ANOVA), Q value = 6.3e-05

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

nPatients Mean (Std.Dev)
ALL 526 57.9 (13.2)
subtype1 126 58.5 (14.3)
subtype2 39 49.3 (10.7)
subtype3 23 58.8 (13.8)
subtype4 100 53.9 (12.6)
subtype5 74 58.6 (12.5)
subtype6 20 60.4 (9.9)
subtype7 104 62.4 (12.3)
subtype8 40 59.8 (13.0)

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

'mRNA CNMF subtypes' versus 'GENDER'

P value = 0.0612 (Chi-square test), Q value = 1

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

nPatients FEMALE MALE
ALL 520 6
subtype1 122 4
subtype2 39 0
subtype3 22 1
subtype4 100 0
subtype5 74 0
subtype6 19 1
subtype7 104 0
subtype8 40 0

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

'mRNA CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.921 (Chi-square test), Q value = 1

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

nPatients NO YES
ALL 148 378
subtype1 33 93
subtype2 11 28
subtype3 5 18
subtype4 34 66
subtype5 21 53
subtype6 5 15
subtype7 28 76
subtype8 11 29

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

'mRNA CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.538 (Chi-square test), Q value = 1

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

nPatients CM0 (I+) M0 M1 MX
ALL 2 496 14 14
subtype1 0 121 1 4
subtype2 1 37 1 0
subtype3 0 21 2 0
subtype4 0 96 3 1
subtype5 0 70 3 1
subtype6 0 19 0 1
subtype7 1 95 3 5
subtype8 0 37 1 2

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

'mRNA CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.06 (Chi-square test), Q value = 1

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

nPatients N0 N0 (I+) N0 (I-) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3C NX
ALL 138 16 101 58 77 20 2 13 32 29 10 18 1 11
subtype1 28 1 23 17 24 5 0 6 7 7 1 3 0 4
subtype2 3 4 7 3 9 2 0 2 4 1 1 2 0 1
subtype3 2 0 5 3 2 1 0 0 5 3 0 1 0 1
subtype4 34 3 26 12 6 4 1 1 5 3 1 3 1 0
subtype5 14 0 10 13 10 2 0 1 6 8 4 4 0 2
subtype6 6 0 5 1 3 0 0 1 1 2 0 0 0 1
subtype7 35 7 20 5 17 4 1 2 4 3 2 3 0 1
subtype8 16 1 5 4 6 2 0 0 0 2 1 2 0 1

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

'mRNA CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.415 (ANOVA), Q value = 1

Table S8.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 406 1.8 (3.5)
subtype1 89 1.6 (2.3)
subtype2 34 1.9 (2.4)
subtype3 17 1.6 (2.3)
subtype4 91 1.5 (3.3)
subtype5 46 2.8 (3.5)
subtype6 15 1.2 (2.5)
subtype7 79 1.8 (4.0)
subtype8 35 2.6 (6.3)

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

'mRNA CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.036 (Chi-square test), Q value = 1

Table S9.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 42 41 6 185 110 77 15 19 14 1 16
subtype1 6 9 4 38 35 21 6 3 1 0 3
subtype2 2 3 0 11 10 8 0 2 1 1 1
subtype3 1 1 0 7 3 6 2 0 2 0 1
subtype4 7 10 0 45 19 9 1 4 3 0 2
subtype5 4 2 0 21 19 15 2 6 3 0 2
subtype6 3 2 0 6 4 4 0 0 0 0 1
subtype7 13 13 0 42 14 9 4 2 3 0 4
subtype8 6 1 2 15 6 5 0 2 1 0 2

Figure S8.  Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #2: 'mRNA cHierClus subtypes'

Table S10.  Get Full Table Description of clustering approach #2: 'mRNA cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 143 131 252
'mRNA cHierClus subtypes' versus 'Time to Death'

P value = 0.609 (logrank test), Q value = 1

Table S11.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 494 65 0.1 - 223.4 (24.2)
subtype1 140 14 0.3 - 157.4 (26.4)
subtype2 120 18 0.1 - 211.5 (23.8)
subtype3 234 33 0.1 - 223.4 (20.2)

Figure S9.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'mRNA cHierClus subtypes' versus 'AGE'

P value = 0.000437 (ANOVA), Q value = 0.032

Table S12.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 526 57.9 (13.2)
subtype1 143 56.4 (12.8)
subtype2 131 55.1 (13.1)
subtype3 252 60.2 (13.2)

Figure S10.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'mRNA cHierClus subtypes' versus 'GENDER'

P value = 0.493 (Fisher's exact test), Q value = 1

Table S13.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 520 6
subtype1 141 2
subtype2 131 0
subtype3 248 4

Figure S11.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'mRNA cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.105 (Fisher's exact test), Q value = 1

Table S14.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 148 378
subtype1 42 101
subtype2 45 86
subtype3 61 191

Figure S12.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'mRNA cHierClus subtypes' versus 'DISTANT.METASTASIS'

P value = 0.173 (Chi-square test), Q value = 1

Table S15.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 496 14 14
subtype1 2 134 2 5
subtype2 0 126 4 1
subtype3 0 236 8 8

Figure S13.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'mRNA cHierClus subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.117 (Chi-square test), Q value = 1

Table S16.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3C NX
ALL 138 16 101 58 77 20 2 13 32 29 10 18 1 11
subtype1 36 9 27 14 24 4 0 1 8 8 3 5 0 4
subtype2 38 3 33 18 10 4 1 1 8 6 3 5 1 0
subtype3 64 4 41 26 43 12 1 11 16 15 4 8 0 7

Figure S14.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'mRNA cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.692 (ANOVA), Q value = 1

Table S17.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 406 1.8 (3.5)
subtype1 125 2.0 (4.1)
subtype2 114 1.7 (3.3)
subtype3 167 1.7 (3.2)

Figure S15.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'mRNA cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0779 (Chi-square test), Q value = 1

Table S18.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 42 41 6 185 110 77 15 19 14 1 16
subtype1 17 12 2 50 24 23 1 5 2 1 6
subtype2 8 12 0 56 25 13 2 8 4 0 3
subtype3 17 17 4 79 61 41 12 6 8 0 7

Figure S16.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

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

Table S19.  Get Full Table Description of clustering approach #3: 'Copy Number Ratio CNMF subtypes'

Cluster Labels 1 2 3 4 5
Number of samples 312 218 67 213 47
'Copy Number Ratio CNMF subtypes' versus 'Time to Death'

P value = 0.376 (logrank test), Q value = 1

Table S20.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 798 94 0.0 - 223.4 (18.1)
subtype1 292 29 0.0 - 223.4 (20.6)
subtype2 201 21 0.1 - 186.4 (17.0)
subtype3 66 8 0.0 - 189.0 (18.1)
subtype4 195 27 0.0 - 211.5 (17.0)
subtype5 44 9 0.7 - 220.9 (20.3)

Figure S17.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.011 (ANOVA), Q value = 0.74

Table S21.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 856 58.5 (13.2)
subtype1 311 57.8 (13.1)
subtype2 218 58.6 (14.2)
subtype3 67 61.9 (11.1)
subtype4 213 57.2 (12.8)
subtype5 47 63.1 (12.2)

Figure S18.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.000381 (Chi-square test), Q value = 0.029

Table S22.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 848 9
subtype1 312 0
subtype2 210 8
subtype3 66 1
subtype4 213 0
subtype5 47 0

Figure S19.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'Copy Number Ratio CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.454 (Chi-square test), Q value = 1

Table S23.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 212 645
subtype1 68 244
subtype2 52 166
subtype3 19 48
subtype4 59 154
subtype5 14 33

Figure S20.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'Copy Number Ratio CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.689 (Chi-square test), Q value = 1

Table S24.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 761 15 79
subtype1 1 272 4 35
subtype2 1 195 5 17
subtype3 0 63 2 2
subtype4 0 190 4 19
subtype5 0 41 0 6

Figure S21.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'Copy Number Ratio CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.364 (Chi-square test), Q value = 1

Table S25.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B N3C NX
ALL 251 22 129 1 102 126 32 2 22 51 53 19 30 2 1 14
subtype1 91 11 57 1 35 44 11 2 8 15 14 7 9 1 0 6
subtype2 52 3 25 0 31 33 7 0 10 18 18 7 9 0 0 5
subtype3 20 1 10 0 5 16 3 0 0 1 4 3 3 0 0 1
subtype4 78 5 32 0 25 23 7 0 3 16 13 1 7 1 1 1
subtype5 10 2 5 0 6 10 4 0 1 1 4 1 2 0 0 1

Figure S22.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'Copy Number Ratio CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.123 (ANOVA), Q value = 1

Table S26.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 714 2.2 (4.4)
subtype1 277 1.9 (3.9)
subtype2 163 2.7 (4.3)
subtype3 52 3.0 (5.7)
subtype4 181 2.0 (4.7)
subtype5 41 2.9 (5.0)

Figure S23.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

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

P value = 0.494 (Chi-square test), Q value = 1

Table S27.  Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 72 64 6 8 291 191 2 124 25 41 15 1 16
subtype1 37 34 2 5 96 66 1 39 7 14 4 0 7
subtype2 12 10 3 1 74 47 1 40 11 12 5 0 2
subtype3 5 2 0 1 23 18 0 7 4 3 2 0 2
subtype4 15 15 1 1 79 50 0 30 3 9 4 1 4
subtype5 3 3 0 0 19 10 0 8 0 3 0 0 1

Figure S24.  Get High-res Image Clustering Approach #3: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #4: 'METHLYATION CNMF'

Table S28.  Get Full Table Description of clustering approach #4: 'METHLYATION CNMF'

Cluster Labels 1 2 3 4 5 6
Number of samples 90 129 87 137 36 80
'METHLYATION CNMF' versus 'Time to Death'

P value = 0.162 (logrank test), Q value = 1

Table S29.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 530 61 0.0 - 223.4 (17.7)
subtype1 85 13 0.2 - 211.5 (19.2)
subtype2 121 10 0.3 - 223.4 (13.7)
subtype3 85 14 0.0 - 173.0 (13.0)
subtype4 130 13 0.1 - 194.3 (19.0)
subtype5 34 3 0.1 - 157.4 (20.4)
subtype6 75 8 0.0 - 130.2 (25.4)

Figure S25.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'AGE'

P value = 0.0015 (ANOVA), Q value = 0.11

Table S30.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 558 57.8 (13.1)
subtype1 90 56.5 (13.2)
subtype2 129 58.3 (12.4)
subtype3 87 63.1 (11.6)
subtype4 137 56.0 (14.3)
subtype5 36 56.2 (13.7)
subtype6 79 56.6 (11.7)

Figure S26.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.0712 (Chi-square test), Q value = 1

Table S31.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 553 6
subtype1 90 0
subtype2 128 1
subtype3 85 2
subtype4 136 1
subtype5 34 2
subtype6 80 0

Figure S27.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

'METHLYATION CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.183 (Chi-square test), Q value = 1

Table S32.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 147 412
subtype1 30 60
subtype2 30 99
subtype3 20 67
subtype4 43 94
subtype5 6 30
subtype6 18 62

Figure S28.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'METHLYATION CNMF' versus 'DISTANT.METASTASIS'

P value = 0.745 (Chi-square test), Q value = 1

Table S33.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 1 481 6 71
subtype1 0 79 2 9
subtype2 1 112 0 16
subtype3 0 72 2 13
subtype4 0 120 1 16
subtype5 0 31 1 4
subtype6 0 67 0 13

Figure S29.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'METHLYATION CNMF' versus 'LYMPH.NODE.METASTASIS'

P value = 0.328 (Chi-square test), Q value = 1

Table S34.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B NX
ALL 152 14 79 1 72 84 23 2 14 38 41 13 17 2 7
subtype1 35 2 15 0 10 10 3 0 0 5 6 1 2 1 0
subtype2 30 3 16 1 13 29 7 1 5 8 6 2 4 0 4
subtype3 23 2 14 0 11 10 4 0 0 8 7 2 4 0 2
subtype4 28 2 17 0 24 23 6 0 5 9 15 2 5 0 1
subtype5 11 1 4 0 3 5 2 0 0 4 1 3 2 0 0
subtype6 25 4 13 0 11 7 1 1 4 4 6 3 0 1 0

Figure S30.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'METHLYATION CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.273 (ANOVA), Q value = 1

Table S35.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 519 2.5 (4.6)
subtype1 86 1.9 (3.3)
subtype2 121 2.3 (4.2)
subtype3 80 3.1 (6.2)
subtype4 123 2.7 (4.3)
subtype5 32 3.6 (5.4)
subtype6 77 1.9 (4.7)

Figure S31.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'METHLYATION CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0398 (Chi-square test), Q value = 1

Table S36.  Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE X
ALL 47 42 2 8 181 126 2 95 15 29 6 5
subtype1 4 6 0 1 37 22 0 12 1 3 2 1
subtype2 11 12 0 2 37 35 0 20 4 6 0 2
subtype3 7 4 1 2 25 20 1 17 0 6 2 2
subtype4 8 8 1 0 41 34 0 30 8 6 1 0
subtype5 2 6 0 1 11 3 0 5 2 5 1 0
subtype6 15 6 0 2 30 12 1 11 0 3 0 0

Figure S32.  Get High-res Image Clustering Approach #4: 'METHLYATION CNMF' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #5: 'RPPA CNMF subtypes'

Table S37.  Get Full Table Description of clustering approach #5: 'RPPA CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 152 136 120
'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.0477 (logrank test), Q value = 1

Table S38.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 380 44 0.1 - 189.0 (24.5)
subtype1 135 22 0.1 - 186.4 (22.8)
subtype2 132 11 0.2 - 146.5 (18.0)
subtype3 113 11 0.3 - 189.0 (28.5)

Figure S33.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'AGE'

P value = 0.0287 (ANOVA), Q value = 1

Table S39.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 408 57.9 (13.1)
subtype1 152 56.1 (13.4)
subtype2 136 60.1 (13.7)
subtype3 120 57.5 (11.7)

Figure S34.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.0564 (Fisher's exact test), Q value = 1

Table S40.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 403 5
subtype1 152 0
subtype2 132 4
subtype3 119 1

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

'RPPA CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.981 (Fisher's exact test), Q value = 1

Table S41.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 125 283
subtype1 47 105
subtype2 42 94
subtype3 36 84

Figure S36.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.323 (Chi-square test), Q value = 1

Table S42.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 1 386 14 7
subtype1 0 143 8 1
subtype2 0 129 3 4
subtype3 1 114 3 2

Figure S37.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'RPPA CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.0979 (Chi-square test), Q value = 1

Table S43.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3C NX
ALL 100 13 79 48 63 8 1 12 27 25 9 15 1 7
subtype1 41 2 27 24 14 3 0 4 11 11 4 7 1 3
subtype2 31 6 28 10 35 4 1 3 7 3 2 4 0 2
subtype3 28 5 24 14 14 1 0 5 9 11 3 4 0 2

Figure S38.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'RPPA CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0305 (ANOVA), Q value = 1

Table S44.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 330 1.9 (3.5)
subtype1 114 2.1 (3.6)
subtype2 113 1.2 (2.0)
subtype3 103 2.4 (4.5)

Figure S39.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'RPPA CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0292 (Chi-square test), Q value = 1

Table S45.  Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE X
ALL 30 31 4 137 94 62 13 15 14 8
subtype1 6 8 0 55 36 24 2 10 8 3
subtype2 11 13 1 46 38 14 6 1 3 3
subtype3 13 10 3 36 20 24 5 4 3 2

Figure S40.  Get High-res Image Clustering Approach #5: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #6: 'RPPA cHierClus subtypes'

Table S46.  Get Full Table Description of clustering approach #6: 'RPPA cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 163 115 130
'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.318 (logrank test), Q value = 1

Table S47.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 380 44 0.1 - 189.0 (24.5)
subtype1 152 17 0.2 - 129.7 (24.2)
subtype2 112 10 0.2 - 173.0 (25.5)
subtype3 116 17 0.1 - 189.0 (21.9)

Figure S41.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'AGE'

P value = 0.00121 (ANOVA), Q value = 0.087

Table S48.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 408 57.9 (13.1)
subtype1 163 60.4 (13.6)
subtype2 115 57.6 (11.8)
subtype3 130 54.8 (13.0)

Figure S42.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.154 (Fisher's exact test), Q value = 1

Table S49.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 403 5
subtype1 159 4
subtype2 114 1
subtype3 130 0

Figure S43.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RPPA cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.917 (Fisher's exact test), Q value = 1

Table S50.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 125 283
subtype1 48 115
subtype2 36 79
subtype3 41 89

Figure S44.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA cHierClus subtypes' versus 'DISTANT.METASTASIS'

P value = 0.665 (Chi-square test), Q value = 1

Table S51.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 1 386 14 7
subtype1 0 153 6 4
subtype2 1 109 3 2
subtype3 0 124 5 1

Figure S45.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'RPPA cHierClus subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.194 (Chi-square test), Q value = 1

Table S52.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3C NX
ALL 100 13 79 48 63 8 1 12 27 25 9 15 1 7
subtype1 42 5 31 14 33 3 1 4 13 6 2 5 0 4
subtype2 22 6 23 11 20 1 0 5 7 11 3 4 0 2
subtype3 36 2 25 23 10 4 0 3 7 8 4 6 1 1

Figure S46.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'RPPA cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0553 (ANOVA), Q value = 1

Table S53.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 330 1.9 (3.5)
subtype1 126 1.3 (2.1)
subtype2 99 2.4 (4.5)
subtype3 105 2.1 (3.7)

Figure S47.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'RPPA cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0496 (Chi-square test), Q value = 1

Table S54.  Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE IIA STAGE IIB STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE X
ALL 30 31 4 137 94 62 13 15 14 8
subtype1 14 12 1 52 42 23 7 2 6 4
subtype2 11 12 3 31 21 23 5 4 3 2
subtype3 5 7 0 54 31 16 1 9 5 2

Figure S48.  Get High-res Image Clustering Approach #6: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #7: 'RNAseq CNMF subtypes'

Table S55.  Get Full Table Description of clustering approach #7: 'RNAseq CNMF subtypes'

Cluster Labels 1 2 3 4 5 6 7 8
Number of samples 150 81 100 131 179 12 151 31
'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.000437 (logrank test), Q value = 0.032

Table S56.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 778 94 0.0 - 223.4 (18.9)
subtype1 143 19 0.1 - 211.5 (19.1)
subtype2 77 7 0.1 - 157.4 (32.0)
subtype3 90 6 0.2 - 223.4 (16.2)
subtype4 119 26 0.0 - 146.5 (20.1)
subtype5 165 19 0.1 - 177.4 (18.2)
subtype6 11 2 0.2 - 130.2 (50.9)
subtype7 145 14 0.3 - 220.9 (15.7)
subtype8 28 1 0.3 - 194.3 (15.3)

Figure S49.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 2.07e-05 (ANOVA), Q value = 0.0016

Table S57.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 834 58.2 (13.1)
subtype1 150 55.7 (12.5)
subtype2 80 53.6 (11.9)
subtype3 100 58.3 (11.9)
subtype4 131 59.7 (13.1)
subtype5 179 57.4 (14.5)
subtype6 12 62.2 (15.8)
subtype7 151 62.2 (12.7)
subtype8 31 58.9 (10.2)

Figure S50.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.0993 (Chi-square test), Q value = 1

Table S58.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 826 9
subtype1 150 0
subtype2 81 0
subtype3 99 1
subtype4 130 1
subtype5 173 6
subtype6 12 0
subtype7 150 1
subtype8 31 0

Figure S51.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.475 (Chi-square test), Q value = 1

Table S59.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 209 626
subtype1 46 104
subtype2 19 62
subtype3 21 79
subtype4 26 105
subtype5 44 135
subtype6 4 8
subtype7 42 109
subtype8 7 24

Figure S52.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.74 (Chi-square test), Q value = 1

Table S60.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 747 15 71
subtype1 0 138 3 9
subtype2 1 71 2 7
subtype3 1 89 1 9
subtype4 0 115 5 11
subtype5 0 163 1 15
subtype6 0 10 0 2
subtype7 0 133 2 16
subtype8 0 28 1 2

Figure S53.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'RNAseq CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.0188 (Chi-square test), Q value = 1

Table S61.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B N3C NX
ALL 240 22 129 1 100 124 32 2 22 50 50 17 29 1 1 15
subtype1 62 4 29 0 16 12 5 1 1 7 6 1 4 1 1 0
subtype2 17 8 13 0 7 14 5 0 3 2 3 3 5 0 0 1
subtype3 30 1 16 0 12 18 1 0 3 7 6 1 3 0 0 2
subtype4 26 1 13 0 21 19 7 0 2 12 12 5 8 0 0 5
subtype5 34 3 30 0 30 27 7 0 7 14 14 3 5 0 0 5
subtype6 4 0 3 0 0 1 0 0 1 1 1 0 1 0 0 0
subtype7 53 5 21 1 9 30 6 1 5 7 6 3 3 0 0 1
subtype8 14 0 4 0 5 3 1 0 0 0 2 1 0 0 0 1

Figure S54.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'RNAseq CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.00172 (ANOVA), Q value = 0.12

Table S62.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 694 2.2 (4.3)
subtype1 137 1.5 (3.1)
subtype2 76 2.8 (6.4)
subtype3 77 1.6 (3.6)
subtype4 90 3.8 (6.0)
subtype5 142 2.1 (3.1)
subtype6 11 3.5 (6.8)
subtype7 133 1.9 (3.5)
subtype8 28 1.5 (3.2)

Figure S55.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'RNAseq CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.00611 (Chi-square test), Q value = 0.42

Table S63.  Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 69 63 7 8 289 186 2 117 23 37 15 1 17
subtype1 11 11 0 1 72 28 0 14 1 6 3 0 2
subtype2 8 5 1 3 24 20 1 9 0 6 2 1 1
subtype3 9 11 0 2 34 20 1 13 3 3 1 0 3
subtype4 8 3 0 1 36 34 0 25 4 10 5 0 5
subtype5 10 10 5 0 54 46 0 33 11 7 1 0 2
subtype6 1 1 0 0 5 2 0 2 0 1 0 0 0
subtype7 18 20 0 0 52 31 0 18 4 3 2 0 3
subtype8 4 2 1 1 12 5 0 3 0 1 1 0 1

Figure S56.  Get High-res Image Clustering Approach #7: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #8: 'RNAseq cHierClus subtypes'

Table S64.  Get Full Table Description of clustering approach #8: 'RNAseq cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 337 209 289
'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0339 (logrank test), Q value = 1

Table S65.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 778 94 0.0 - 223.4 (18.9)
subtype1 306 42 0.0 - 223.4 (17.0)
subtype2 193 30 0.0 - 211.5 (19.4)
subtype3 279 22 0.1 - 194.3 (23.4)

Figure S57.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 3.71e-05 (ANOVA), Q value = 0.0029

Table S66.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 834 58.2 (13.1)
subtype1 337 60.7 (13.4)
subtype2 209 56.9 (12.9)
subtype3 288 56.2 (12.6)

Figure S58.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.0115 (Fisher's exact test), Q value = 0.76

Table S67.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 826 9
subtype1 329 8
subtype2 209 0
subtype3 288 1

Figure S59.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.239 (Fisher's exact test), Q value = 1

Table S68.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 209 626
subtype1 74 263
subtype2 57 152
subtype3 78 211

Figure S60.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'DISTANT.METASTASIS'

P value = 0.3 (Chi-square test), Q value = 1

Table S69.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 747 15 71
subtype1 0 300 7 30
subtype2 0 193 4 12
subtype3 2 254 4 29

Figure S61.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'RNAseq cHierClus subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.041 (Chi-square test), Q value = 1

Table S70.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B N3C NX
ALL 240 22 129 1 100 124 32 2 22 50 50 17 29 1 1 15
subtype1 91 4 37 0 47 60 15 0 10 25 22 7 12 0 0 7
subtype2 71 5 39 0 27 17 8 1 2 12 11 4 7 1 1 3
subtype3 78 13 53 1 26 47 9 1 10 13 17 6 10 0 0 5

Figure S62.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'RNAseq cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.919 (ANOVA), Q value = 1

Table S71.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 694 2.2 (4.3)
subtype1 253 2.3 (3.6)
subtype2 177 2.1 (4.7)
subtype3 264 2.1 (4.6)

Figure S63.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'RNAseq cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 1.43e-05 (Chi-square test), Q value = 0.0011

Table S72.  Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 69 63 7 8 289 186 2 117 23 37 15 1 17
subtype1 14 19 4 2 113 84 1 54 20 12 7 0 7
subtype2 15 15 0 1 91 42 1 22 2 11 4 0 4
subtype3 40 29 3 5 85 60 0 41 1 14 4 1 6

Figure S64.  Get High-res Image Clustering Approach #8: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #9: 'MIRSEQ CNMF'

Table S73.  Get Full Table Description of clustering approach #9: 'MIRSEQ CNMF'

Cluster Labels 1 2 3
Number of samples 174 431 243
'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0226 (logrank test), Q value = 1

Table S74.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 789 95 0.0 - 223.4 (18.5)
subtype1 167 20 0.4 - 159.1 (25.7)
subtype2 400 39 0.0 - 223.4 (17.9)
subtype3 222 36 0.0 - 211.5 (18.0)

Figure S65.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 0.0567 (ANOVA), Q value = 1

Table S75.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 847 58.4 (13.2)
subtype1 173 56.9 (12.8)
subtype2 431 59.5 (13.4)
subtype3 243 57.6 (13.1)

Figure S66.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.327 (Fisher's exact test), Q value = 1

Table S76.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 839 9
subtype1 174 0
subtype2 425 6
subtype3 240 3

Figure S67.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #3: 'GENDER'

'MIRSEQ CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.243 (Fisher's exact test), Q value = 1

Table S77.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 210 638
subtype1 35 139
subtype2 115 316
subtype3 60 183

Figure S68.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CNMF' versus 'DISTANT.METASTASIS'

P value = 0.697 (Chi-square test), Q value = 1

Table S78.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 756 14 76
subtype1 1 155 1 17
subtype2 1 386 7 37
subtype3 0 215 6 22

Figure S69.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'MIRSEQ CNMF' versus 'LYMPH.NODE.METASTASIS'

P value = 0.00385 (Chi-square test), Q value = 0.27

Table S79.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B N3C NX
ALL 248 22 128 1 101 126 32 2 23 50 50 18 29 2 1 15
subtype1 48 10 27 1 18 25 7 1 9 2 11 6 8 1 0 0
subtype2 119 9 62 0 49 80 17 1 10 30 23 8 12 0 0 11
subtype3 81 3 39 0 34 21 8 0 4 18 16 4 9 1 1 4

Figure S70.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'MIRSEQ CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.141 (ANOVA), Q value = 1

Table S80.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 709 2.2 (4.4)
subtype1 163 2.8 (5.9)
subtype2 348 1.9 (3.4)
subtype3 198 2.2 (4.4)

Figure S71.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 4.74e-05 (Chi-square test), Q value = 0.0036

Table S81.  Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 72 64 7 8 289 189 2 120 24 40 14 1 17
subtype1 26 15 1 6 52 36 1 18 4 14 1 0 0
subtype2 36 36 5 1 137 103 1 64 16 13 7 0 12
subtype3 10 13 1 1 100 50 0 38 4 13 6 1 5

Figure S72.  Get High-res Image Clustering Approach #9: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Clustering Approach #10: 'MIRSEQ CHIERARCHICAL'

Table S82.  Get Full Table Description of clustering approach #10: 'MIRSEQ CHIERARCHICAL'

Cluster Labels 1 2 3
Number of samples 195 594 59
'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.33 (logrank test), Q value = 1

Table S83.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 789 95 0.0 - 223.4 (18.5)
subtype1 182 28 0.1 - 211.5 (20.1)
subtype2 556 61 0.1 - 223.4 (18.2)
subtype3 51 6 0.0 - 113.8 (17.0)

Figure S73.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.0651 (ANOVA), Q value = 1

Table S84.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 847 58.4 (13.2)
subtype1 195 56.7 (12.6)
subtype2 593 59.1 (13.5)
subtype3 59 57.2 (12.3)

Figure S74.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.187 (Fisher's exact test), Q value = 1

Table S85.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 839 9
subtype1 194 1
subtype2 588 6
subtype3 57 2

Figure S75.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.285 (Fisher's exact test), Q value = 1

Table S86.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 210 638
subtype1 53 142
subtype2 147 447
subtype3 10 49

Figure S76.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CHIERARCHICAL' versus 'DISTANT.METASTASIS'

P value = 0.554 (Chi-square test), Q value = 1

Table S87.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'DISTANT.METASTASIS'

nPatients CM0 (I+) M0 M1 MX
ALL 2 756 14 76
subtype1 0 179 4 12
subtype2 2 526 8 58
subtype3 0 51 2 6

Figure S77.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'DISTANT.METASTASIS'

'MIRSEQ CHIERARCHICAL' versus 'LYMPH.NODE.METASTASIS'

P value = 0.51 (Chi-square test), Q value = 1

Table S88.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

nPatients N0 N0 (I+) N0 (I-) N0 (MOL+) N1 N1A N1B N1C N1MI N2 N2A N3 N3A N3B N3C NX
ALL 248 22 128 1 101 126 32 2 23 50 50 18 29 2 1 15
subtype1 66 5 36 0 24 18 7 1 1 11 11 2 8 1 1 3
subtype2 160 17 84 1 70 101 23 1 21 33 36 15 20 1 0 11
subtype3 22 0 8 0 7 7 2 0 1 6 3 1 1 0 0 1

Figure S78.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'LYMPH.NODE.METASTASIS'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.804 (ANOVA), Q value = 1

Table S89.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

nPatients Mean (Std.Dev)
ALL 709 2.2 (4.4)
subtype1 168 2.0 (4.6)
subtype2 496 2.3 (4.3)
subtype3 45 2.2 (4.8)

Figure S79.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.593 (Chi-square test), Q value = 1

Table S90.  Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

nPatients STAGE I STAGE IA STAGE IB STAGE II STAGE IIA STAGE IIB STAGE III STAGE IIIA STAGE IIIB STAGE IIIC STAGE IV STAGE TIS STAGE X
ALL 72 64 7 8 289 189 2 120 24 40 14 1 17
subtype1 13 15 1 1 81 38 0 23 3 11 4 0 4
subtype2 57 44 6 7 184 137 2 87 21 28 8 1 12
subtype3 2 5 0 0 24 14 0 10 0 1 2 0 1

Figure S80.  Get High-res Image Clustering Approach #10: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NEOPLASM.DISEASESTAGE'

Methods & Data
Input

  • Cluster data file = BRCA-TP.mergedcluster.txt

  • Clinical data file = BRCA-TP.clin.merged.picked.txt

  • Number of patients = 873

  • Number of clustering approaches = 10

  • Number of selected clinical features = 9

  • 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

ANOVA analysis

For continuous numerical clinical features, one-way analysis of variance (Howell 2002) was applied to compare the clinical values between tumor subtypes using 'anova' function in R

Chi-square test

For multi-class clinical features (nominal or ordinal), Chi-square tests (Greenwood and Nikulin 1996) were used to estimate the P values using the 'chisq.test' 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

This is an experimental feature. The full results of the analysis summarized in this report can be downloaded from the TCGA Data Coordination Center.

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] Howell, D, Statistical Methods for Psychology. (5th ed.), Duxbury Press:324-5 (2002)
[5] Greenwood and Nikulin, A guide to chi-squared testing, Wiley, New York. ISBN 047155779X (1996)
[6] 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)
[7] 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 © 2013 Broad Institute TCGA GDAC as part of the TCGA Research Network. All rights reserved.
Made with Nozzle