Correlation between molecular cancer subtypes and selected clinical features

Overview

Introduction

This pipeline computes the correlation between cancer subtypes identified by different molecular patterns and selected clinical features.

Summary

Testing the association between subtypes identified by 5 different clustering approaches and 4 clinical features across 66 patients, 2 significant findings detected with P value < 0.05.

3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes do not correlate to any clinical features.
CNMF clustering analysis on sequencing-based mRNA expression data identified 2 subtypes that do not correlate to any clinical features.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'Time to Death'.
CNMF clustering analysis on sequencing-based miR expression data identified 5 subtypes that correlate to 'Time to Death'.
Consensus hierarchical clustering analysis on sequencing-based miR expression data identified 2 subtypes that 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 5 different clustering approaches and 4 clinical features. Shown in the table are P values from statistical tests. Thresholded by P value < 0.05, 2 significant findings detected.


Clinical Features	Time to Death	AGE	GENDER	KARNOFSKY PERFORMANCE SCORE
Statistical Tests	logrank test	t-test	Fisher's exact test	t-test
METHLYATION CNMF	0.0846	0.754	0.187	0.713
RNAseq CNMF subtypes	0.0614	0.966	0.267
RNAseq cHierClus subtypes	0.0098	0.929	0.189
MIRseq CNMF subtypes	0.00309	0.52	0.927	0.573
MIRseq cHierClus subtypes	0.5	0.107	1

Clustering Approach #1: 'METHLYATION CNMF'

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

Cluster Labels	1	2	3
Number of samples	18	26	21

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.0846 (logrank test)

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	65	24	0.4 - 118.9 (7.8)
subtype1	18	5	1.5 - 118.9 (8.7)
subtype2	26	12	0.4 - 75.3 (8.3)
subtype3	21	7	0.5 - 26.1 (7.0)

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

'METHLYATION CNMF' versus 'AGE'

P value = 0.754 (ANOVA)

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

	nPatients	Mean (Std.Dev)
ALL	65	68.0 (10.5)
subtype1	18	66.4 (10.3)
subtype2	26	68.8 (10.6)
subtype3	21	68.2 (10.7)

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

'METHLYATION CNMF' versus 'GENDER'

P value = 0.187 (Fisher's exact test)

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

nPatients	FEMALE	MALE
ALL	42	23
subtype1	10	8
subtype2	15	11
subtype3	17	4

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

'METHLYATION CNMF' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.713 (ANOVA)

Table S5. Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

	nPatients	Mean (Std.Dev)
ALL	16	79.4 (17.7)
subtype1	4	85.0 (5.8)
subtype2	3	73.3 (28.9)
subtype3	9	78.9 (18.3)

Figure S4. Get High-res Image Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

Clustering Approach #2: 'RNAseq CNMF subtypes'

Table S6. Get Full Table Description of clustering approach #2: 'RNAseq CNMF subtypes'

Cluster Labels	1	2
Number of samples	18	14

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.0614 (logrank test)

Table S7. Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	32	15	0.4 - 118.9 (7.8)
subtype1	18	6	0.5 - 118.9 (7.8)
subtype2	14	9	0.4 - 26.1 (8.0)

Figure S5. Get High-res Image Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.966 (t-test)

Table S8. Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	32	70.3 (8.9)
subtype1	18	70.2 (9.6)
subtype2	14	70.4 (8.3)

Figure S6. Get High-res Image Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.267 (Fisher's exact test)

Table S9. Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	22	10
subtype1	14	4
subtype2	8	6

Figure S7. Get High-res Image Clustering Approach #2: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

Clustering Approach #3: 'RNAseq cHierClus subtypes'

Table S10. Get Full Table Description of clustering approach #3: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	12	8	12

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0098 (logrank test)

Table S11. Clustering Approach #3: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	32	15	0.4 - 118.9 (7.8)
subtype1	12	7	0.4 - 19.0 (8.0)
subtype2	8	5	0.5 - 19.5 (6.8)
subtype3	12	3	2.0 - 118.9 (8.7)

Figure S8. Get High-res Image Clustering Approach #3: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.929 (ANOVA)

Table S12. Clustering Approach #3: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	32	70.3 (8.9)
subtype1	12	69.5 (8.5)
subtype2	8	70.5 (10.1)
subtype3	12	70.9 (9.2)

Figure S9. Get High-res Image Clustering Approach #3: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.189 (Fisher's exact test)

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

nPatients	FEMALE	MALE
ALL	22	10
subtype1	6	6
subtype2	7	1
subtype3	9	3

Figure S10. Get High-res Image Clustering Approach #3: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

Clustering Approach #4: 'MIRseq CNMF subtypes'

Table S14. Get Full Table Description of clustering approach #4: 'MIRseq CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	12	29	10	3	12

'MIRseq CNMF subtypes' versus 'Time to Death'

P value = 0.00309 (logrank test)

Table S15. Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	66	25	0.4 - 118.9 (7.7)
subtype1	12	4	1.5 - 118.9 (9.5)
subtype2	29	14	0.4 - 49.2 (8.2)
subtype3	10	3	0.7 - 36.4 (10.4)
subtype4	3	2	5.1 - 6.6 (6.2)
subtype5	12	2	0.5 - 100.5 (6.8)

Figure S11. Get High-res Image Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq CNMF subtypes' versus 'AGE'

P value = 0.52 (ANOVA)

Table S16. Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	66	67.9 (10.4)
subtype1	12	70.6 (8.5)
subtype2	29	67.7 (10.5)
subtype3	10	63.1 (11.7)
subtype4	3	71.0 (10.6)
subtype5	12	68.8 (11.0)

Figure S12. Get High-res Image Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq CNMF subtypes' versus 'GENDER'

P value = 0.927 (Chi-square test)

Table S17. Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	43	23
subtype1	9	3
subtype2	19	10
subtype3	6	4
subtype4	2	1
subtype5	7	5

Figure S13. Get High-res Image Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq CNMF subtypes' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.573 (ANOVA)

Table S18. Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

	nPatients	Mean (Std.Dev)
ALL	16	79.4 (17.7)
subtype2	8	76.2 (22.6)
subtype3	5	82.0 (13.0)
subtype4	1	90.0 (NA)
subtype5	2	80.0 (14.1)

Figure S14. Get High-res Image Clustering Approach #4: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

Clustering Approach #5: 'MIRseq cHierClus subtypes'

Table S19. Get Full Table Description of clustering approach #5: 'MIRseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	2	12	52

'MIRseq cHierClus subtypes' versus 'Time to Death'

P value = 0.5 (logrank test)

Table S20. Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	64	24	0.4 - 118.9 (8.0)
subtype2	12	3	1.5 - 118.9 (7.8)
subtype3	52	21	0.4 - 100.5 (8.0)

Figure S15. Get High-res Image Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq cHierClus subtypes' versus 'AGE'

P value = 0.107 (t-test)

Table S21. Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	64	68.1 (10.5)
subtype2	12	72.0 (8.5)
subtype3	52	67.2 (10.8)

Figure S16. Get High-res Image Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq cHierClus subtypes' versus 'GENDER'

P value = 1 (Fisher's exact test)

Table S22. Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	41	23
subtype2	8	4
subtype3	33	19

Figure S17. Get High-res Image Clustering Approach #5: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

Methods & Data

Input

Cluster data file = BLCA.mergedcluster.txt
Clinical data file = BLCA.clin.merged.picked.txt
Number of patients = 66
Number of clustering approaches = 5
Number of selected clinical features = 4
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

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

Student's t-test analysis

For continuous numerical clinical features, two-tailed Student's t test with unequal variance (Lehmann and Romano 2005) was applied to compare the clinical values between two tumor subtypes using 't.test' 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

References

[1] Brunet et al., Metagenes and molecular pattern discovery using matrix factorization, PNAS 101(12):4164-9 (2004)

[2] Monti et al., Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data, Machine Learning 52(1):91-118 (2003)

[3] Bland and Altman, Statistics notes: The logrank test, BMJ 328(7447):1073 (2004)

[4] Howell, D, Statistical Methods for Psychology. (5th ed.), Duxbury Press:324-5 (2002)

[5] 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)

[6] Lehmann and Romano, Testing Statistical Hypotheses (3E ed.), New York: Springer. ISBN 0387988645 (2005)

[7] Greenwood and Nikulin, A guide to chi-squared testing, Wiley, New York. ISBN 047155779X (1996)

Made with Nozzle