Correlation between copy number variations of arm-level result and selected clinical features

Overview

Introduction

This pipeline computes the correlation between significant arm-level copy number variations (cnvs) and selected clinical features.

Summary

Testing the association between subtypes identified by 22 different clustering approaches and 3 clinical features across 191 patients, 4 significant findings detected with Q value < 0.25.

2 subtypes identified in current cancer cohort by '1p gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '4p gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '4q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '8p gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '8q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '10q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '11p gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '11q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '13q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '17q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '19p gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '19q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '21q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '22q gain mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '5q loss mutation analysis'. These subtypes correlate to 'Time to Death'.
2 subtypes identified in current cancer cohort by '7p loss mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '7q loss mutation analysis'. These subtypes correlate to 'Time to Death'.
2 subtypes identified in current cancer cohort by '12p loss mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '17p loss mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '17q loss mutation analysis'. These subtypes do not correlate to any clinical features.
2 subtypes identified in current cancer cohort by '18p loss mutation analysis'. These subtypes correlate to 'Time to Death'.
2 subtypes identified in current cancer cohort by '18q loss mutation analysis'. These subtypes correlate to 'Time to Death'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 22 different clustering approaches and 3 clinical features. Shown in the table are P values (Q values). Thresholded by Q value < 0.25, 4 significant findings detected.


Clinical Features	Time to Death	AGE	GENDER
Statistical Tests	logrank test	t-test	Fisher's exact test
1p gain		0.0913 (1.00)	0.592 (1.00)
4p gain	0.587 (1.00)	0.425 (1.00)	0.627 (1.00)
4q gain	0.587 (1.00)	0.425 (1.00)	0.627 (1.00)
8p gain	0.48 (1.00)	0.119 (1.00)	0.0597 (1.00)
8q gain	0.529 (1.00)	0.168 (1.00)	0.0382 (1.00)
10q gain		0.947 (1.00)	0.252 (1.00)
11p gain		0.188 (1.00)	1 (1.00)
11q gain	0.57 (1.00)	0.0154 (0.834)	1 (1.00)
13q gain		0.511 (1.00)	0.592 (1.00)
17q gain	0.729 (1.00)	0.29 (1.00)	1 (1.00)
19p gain	0.347 (1.00)	0.789 (1.00)	0.0642 (1.00)
19q gain	0.347 (1.00)	0.789 (1.00)	0.0642 (1.00)
21q gain	0.0186 (0.966)	0.616 (1.00)	0.0732 (1.00)
22q gain	0.887 (1.00)	0.00598 (0.341)	0.73 (1.00)
5q loss	0.00073 (0.0438)	0.0601 (1.00)	0.0325 (1.00)
7p loss	0.00716 (0.401)	0.325 (1.00)	1 (1.00)
7q loss	0.00315 (0.183)	0.325 (1.00)	0.805 (1.00)
12p loss		0.514 (1.00)	1 (1.00)
17p loss	0.376 (1.00)	0.308 (1.00)	0.114 (1.00)
17q loss	0.0166 (0.882)	0.0075 (0.412)	0.378 (1.00)
18p loss	0.00202 (0.119)	0.355 (1.00)	0.0642 (1.00)
18q loss	0.000113 (0.00689)	0.0297 (1.00)	0.252 (1.00)

Clustering Approach #1: '1p gain mutation analysis'

Table S1. Get Full Table Description of clustering approach #1: '1p gain mutation analysis'

Cluster Labels	1P GAIN MUTATED	1P GAIN WILD-TYPE
Number of samples	3	188

Clustering Approach #2: '4p gain mutation analysis'

Table S2. Get Full Table Description of clustering approach #2: '4p gain mutation analysis'

Cluster Labels	4P GAIN MUTATED	4P GAIN WILD-TYPE
Number of samples	4	187

Clustering Approach #3: '4q gain mutation analysis'

Table S3. Get Full Table Description of clustering approach #3: '4q gain mutation analysis'

Cluster Labels	4Q GAIN MUTATED	4Q GAIN WILD-TYPE
Number of samples	4	187

Clustering Approach #4: '8p gain mutation analysis'

Table S4. Get Full Table Description of clustering approach #4: '8p gain mutation analysis'

Cluster Labels	8P GAIN MUTATED	8P GAIN WILD-TYPE
Number of samples	20	171

Clustering Approach #5: '8q gain mutation analysis'

Table S5. Get Full Table Description of clustering approach #5: '8q gain mutation analysis'

Cluster Labels	8Q GAIN MUTATED	8Q GAIN WILD-TYPE
Number of samples	21	170

Clustering Approach #6: '10q gain mutation analysis'

Table S6. Get Full Table Description of clustering approach #6: '10q gain mutation analysis'

Cluster Labels	10Q GAIN MUTATED	10Q GAIN WILD-TYPE
Number of samples	3	188

Clustering Approach #7: '11p gain mutation analysis'

Table S7. Get Full Table Description of clustering approach #7: '11p gain mutation analysis'

Cluster Labels	11P GAIN MUTATED	11P GAIN WILD-TYPE
Number of samples	4	187

Clustering Approach #8: '11q gain mutation analysis'

Table S8. Get Full Table Description of clustering approach #8: '11q gain mutation analysis'

Cluster Labels	11Q GAIN MUTATED	11Q GAIN WILD-TYPE
Number of samples	7	184

Clustering Approach #9: '13q gain mutation analysis'

Table S9. Get Full Table Description of clustering approach #9: '13q gain mutation analysis'

Cluster Labels	13Q GAIN MUTATED	13Q GAIN WILD-TYPE
Number of samples	3	188

Clustering Approach #10: '17q gain mutation analysis'

Table S10. Get Full Table Description of clustering approach #10: '17q gain mutation analysis'

Cluster Labels	17Q GAIN MUTATED	17Q GAIN WILD-TYPE
Number of samples	3	188

Clustering Approach #11: '19p gain mutation analysis'

Table S11. Get Full Table Description of clustering approach #11: '19p gain mutation analysis'

Cluster Labels	19P GAIN MUTATED	19P GAIN WILD-TYPE
Number of samples	5	186

Clustering Approach #12: '19q gain mutation analysis'

Table S12. Get Full Table Description of clustering approach #12: '19q gain mutation analysis'

Cluster Labels	19Q GAIN MUTATED	19Q GAIN WILD-TYPE
Number of samples	5	186

Clustering Approach #13: '21q gain mutation analysis'

Table S13. Get Full Table Description of clustering approach #13: '21q gain mutation analysis'

Cluster Labels	21Q GAIN MUTATED	21Q GAIN WILD-TYPE
Number of samples	8	183

Clustering Approach #14: '22q gain mutation analysis'

Table S14. Get Full Table Description of clustering approach #14: '22q gain mutation analysis'

Cluster Labels	22Q GAIN MUTATED	22Q GAIN WILD-TYPE
Number of samples	8	183

Clustering Approach #15: '5q loss mutation analysis'

Table S15. Get Full Table Description of clustering approach #15: '5q loss mutation analysis'

Cluster Labels	5Q LOSS MUTATED	5Q LOSS WILD-TYPE
Number of samples	6	185

'5q loss mutation analysis' versus 'Time to Death'

P value = 0.00073 (logrank test), Q value = 0.044

Table S16. Clustering Approach #15: '5q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	168	106	0.9 - 94.1 (12.0)
5Q LOSS MUTATED	6	6	1.0 - 12.0 (7.0)
5Q LOSS WILD-TYPE	162	100	0.9 - 94.1 (12.5)

Figure S1. Get High-res Image Clustering Approach #15: '5q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

Clustering Approach #16: '7p loss mutation analysis'

Table S17. Get Full Table Description of clustering approach #16: '7p loss mutation analysis'

Cluster Labels	7P LOSS MUTATED	7P LOSS WILD-TYPE
Number of samples	15	176

Clustering Approach #17: '7q loss mutation analysis'

Table S18. Get Full Table Description of clustering approach #17: '7q loss mutation analysis'

Cluster Labels	7Q LOSS MUTATED	7Q LOSS WILD-TYPE
Number of samples	18	173

'7q loss mutation analysis' versus 'Time to Death'

P value = 0.00315 (logrank test), Q value = 0.18

Table S19. Clustering Approach #17: '7q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	168	106	0.9 - 94.1 (12.0)
7Q LOSS MUTATED	15	13	1.0 - 40.0 (9.0)
7Q LOSS WILD-TYPE	153	93	0.9 - 94.1 (13.9)

Figure S2. Get High-res Image Clustering Approach #17: '7q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

Clustering Approach #18: '12p loss mutation analysis'

Table S20. Get Full Table Description of clustering approach #18: '12p loss mutation analysis'

Cluster Labels	12P LOSS MUTATED	12P LOSS WILD-TYPE
Number of samples	3	188

Clustering Approach #19: '17p loss mutation analysis'

Table S21. Get Full Table Description of clustering approach #19: '17p loss mutation analysis'

Cluster Labels	17P LOSS MUTATED	17P LOSS WILD-TYPE
Number of samples	10	181

Clustering Approach #20: '17q loss mutation analysis'

Table S22. Get Full Table Description of clustering approach #20: '17q loss mutation analysis'

Cluster Labels	17Q LOSS MUTATED	17Q LOSS WILD-TYPE
Number of samples	5	186

Clustering Approach #21: '18p loss mutation analysis'

Table S23. Get Full Table Description of clustering approach #21: '18p loss mutation analysis'

Cluster Labels	18P LOSS MUTATED	18P LOSS WILD-TYPE
Number of samples	5	186

'18p loss mutation analysis' versus 'Time to Death'

P value = 0.00202 (logrank test), Q value = 0.12

Table S24. Clustering Approach #21: '18p loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	168	106	0.9 - 94.1 (12.0)
18P LOSS MUTATED	5	5	1.0 - 12.0 (7.0)
18P LOSS WILD-TYPE	163	101	0.9 - 94.1 (12.0)

Figure S3. Get High-res Image Clustering Approach #21: '18p loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

Clustering Approach #22: '18q loss mutation analysis'

Table S25. Get Full Table Description of clustering approach #22: '18q loss mutation analysis'

Cluster Labels	18Q LOSS MUTATED	18Q LOSS WILD-TYPE
Number of samples	3	188

'18q loss mutation analysis' versus 'Time to Death'

P value = 0.000113 (logrank test), Q value = 0.0069

Table S26. Clustering Approach #22: '18q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	168	106	0.9 - 94.1 (12.0)
18Q LOSS MUTATED	3	3	1.0 - 7.0 (2.0)
18Q LOSS WILD-TYPE	165	103	0.9 - 94.1 (12.0)

Figure S4. Get High-res Image Clustering Approach #22: '18q loss mutation analysis' versus Clinical Feature #1: 'Time to Death'

Methods & Data

Input

Cluster data file = broad_values_by_arm.mutsig.cluster.txt
Clinical data file = LAML-TB.clin.merged.picked.txt
Number of patients = 191
Number of clustering approaches = 22
Number of selected clinical features = 3
Exclude small clusters that include fewer than K patients, K = 3

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

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

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.

References

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

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

[3] 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)

[4] Benjamini and Hochberg, Controlling the false discovery rate: a practical and powerful approach to multiple testing, Journal of the Royal Statistical Society Series B 59:289-300 (1995)

Made with Nozzle