Correlation between gene mutation status and molecular subtypes

Overview

Introduction

This pipeline computes the correlation between significantly recurrent gene mutations and molecular subtypes.

Summary

Testing the association between mutation status of 16 genes and 7 molecular subtypes across 69 patients, 3 significant findings detected with P value < 0.05 and Q value < 0.25.

KRAS mutation correlated to 'CN_CNMF'.
TP53 mutation correlated to 'MRNA_CNMF' and 'MRNA_CHIERARCHICAL'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between mutation status of 16 genes and 7 molecular subtypes. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 3 significant findings detected.


		Clinical Features	MRNA CNMF	MRNA CHIERARCHICAL	CN CNMF	RPPA CNMF	RPPA CHIERARCHICAL	MIRSEQ CNMF	MIRSEQ CHIERARCHICAL
	nMutated (%)	nWild-Type	Fisher's exact test	Fisher's exact test	Chi-square test	Fisher's exact test	Chi-square test	Fisher's exact test	Fisher's exact test
TP53	45 (65%)	24	0.000229 (0.0238)	0.00135 (0.139)	0.21 (1.00)	0.16 (1.00)	0.603 (1.00)	0.812 (1.00)	0.288 (1.00)
KRAS	38 (55%)	31	0.342 (1.00)	0.18 (1.00)	0.00233 (0.238)	0.83 (1.00)	0.817 (1.00)	0.292 (1.00)	0.0465 (1.00)
APC	57 (83%)	12	0.85 (1.00)	0.78 (1.00)	0.0561 (1.00)	0.362 (1.00)	0.598 (1.00)	0.337 (1.00)	1 (1.00)
SMAD4	8 (12%)	61	0.105 (1.00)	0.361 (1.00)	0.914 (1.00)	0.152 (1.00)	0.18 (1.00)	0.309 (1.00)	1 (1.00)
KIAA1804	9 (13%)	60	0.322 (1.00)	0.578 (1.00)	0.0997 (1.00)	1 (1.00)	0.82 (1.00)	0.805 (1.00)	1 (1.00)
FBXW7	9 (13%)	60	0.148 (1.00)	0.0962 (1.00)	0.409 (1.00)	0.688 (1.00)	0.585 (1.00)	0.805 (1.00)	1 (1.00)
NRAS	5 (7%)	64	0.856 (1.00)	1 (1.00)	0.675 (1.00)			0.0969 (1.00)	1 (1.00)
TCF7L2	7 (10%)	62	0.883 (1.00)	0.683 (1.00)	0.709 (1.00)	0.152 (1.00)	0.528 (1.00)	1 (1.00)	1 (1.00)
PIK3CA	7 (10%)	62	1 (1.00)	0.731 (1.00)	0.814 (1.00)	0.465 (1.00)	0.488 (1.00)	0.343 (1.00)	0.328 (1.00)
OPCML	6 (9%)	63	0.228 (1.00)	0.0972 (1.00)	0.508 (1.00)	0.675 (1.00)	0.13 (1.00)	1 (1.00)	1 (1.00)
SMAD2	5 (7%)	64	1 (1.00)	1 (1.00)	0.29 (1.00)	0.607 (1.00)	0.0281 (1.00)	1 (1.00)	1 (1.00)
SPATA8	3 (4%)	66			0.388 (1.00)			0.432 (1.00)	0.176 (1.00)
ERBB2	4 (6%)	65	0.54 (1.00)	1 (1.00)	0.576 (1.00)	0.675 (1.00)	0.13 (1.00)	1 (1.00)	1 (1.00)
IL1RAPL2	5 (7%)	64	0.613 (1.00)	0.519 (1.00)	0.785 (1.00)			0.273 (1.00)	1 (1.00)
FAM123B	6 (9%)	63	1 (1.00)	0.731 (1.00)	0.463 (1.00)	0.159 (1.00)	0.906 (1.00)	1 (1.00)	1 (1.00)
ZIM3	5 (7%)	64	0.827 (1.00)	0.668 (1.00)	0.433 (1.00)	0.36 (1.00)	0.337 (1.00)	0.387 (1.00)	0.28 (1.00)

'KRAS MUTATION STATUS' versus 'CN_CNMF'

P value = 0.00233 (Chi-square test), Q value = 0.24

Table S1. Gene #2: 'KRAS MUTATION STATUS' versus Clinical Feature #3: 'CN_CNMF'

nPatients	CLUS_1	CLUS_2	CLUS_3	CLUS_4	CLUS_5
ALL	1	15	17	25	10
KRAS MUTATED	1	3	7	19	8
KRAS WILD-TYPE	0	12	10	6	2

Figure S1. Get High-res Image Gene #2: 'KRAS MUTATION STATUS' versus Clinical Feature #3: 'CN_CNMF'

'TP53 MUTATION STATUS' versus 'MRNA_CNMF'

P value = 0.000229 (Fisher's exact test), Q value = 0.024

Table S2. Gene #3: 'TP53 MUTATION STATUS' versus Clinical Feature #1: 'MRNA_CNMF'

nPatients	CLUS_1	CLUS_2	CLUS_3
ALL	24	19	21
TP53 MUTATED	8	13	19
TP53 WILD-TYPE	16	6	2

Figure S2. Get High-res Image Gene #3: 'TP53 MUTATION STATUS' versus Clinical Feature #1: 'MRNA_CNMF'

'TP53 MUTATION STATUS' versus 'MRNA_CHIERARCHICAL'

P value = 0.00135 (Fisher's exact test), Q value = 0.14

Table S3. Gene #3: 'TP53 MUTATION STATUS' versus Clinical Feature #2: 'MRNA_CHIERARCHICAL'

nPatients	CLUS_1	CLUS_2	CLUS_3
ALL	17	23	24
TP53 MUTATED	15	8	17
TP53 WILD-TYPE	2	15	7

Figure S3. Get High-res Image Gene #3: 'TP53 MUTATION STATUS' versus Clinical Feature #2: 'MRNA_CHIERARCHICAL'

Methods & Data

Input

Mutation data file = READ-TP.mutsig.cluster.txt
Molecular subtypes file = READ-TP.transferedmergedcluster.txt
Number of patients = 69
Number of significantly mutated genes = 16
Number of Molecular subtypes = 7
Exclude genes that fewer than K tumors have mutations, K = 3

Fisher's exact test

For binary or multi-class clinical features (nominal or ordinal), two-tailed Fisher's exact tests (Fisher 1922) were used to estimate the P values using the 'fisher.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

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] 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)

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

[3] 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)