Differentially mutated subnetworks discovery

Morteza Chalabi Hajkarim,Fabio Vandin,Eli Upfal

doi:10.1186/s13015-019-0146-7

Abstract

ProblemWe study the problem of identifying differentially mutated subnetworks of a large gene–gene interaction network, that is, subnetworks that display a significant difference in mutation frequency in two sets of cancer samples. We formally define the associated computational problem and show that the problem is NP-hard.AlgorithmWe propose a novel and efficient algorithm, called DAMOKLE, to identify differentially mutated subnetworks given genome-wide mutation data for two sets of cancer samples. We prove that DAMOKLE identifies subnetworks with statistically significant difference in mutation frequency when the data comes from a reasonable generative model, provided enough samples are available.Experimental resultsWe test DAMOKLE on simulated and real data, showing that DAMOKLE does indeed find subnetworks with significant differences in mutation frequency and that it provides novel insights into the molecular mechanisms of the disease not revealed by standard methods.

Highlights

The analysis of molecular measurements from large collections of cancer samples has revolutionized our understanding of the processes leading to a tumour through somatic mutations, changes of the DNA appearing during the lifetime of an individual [1]
Experimental results: We test DifferentiAlly Mutated subnetwOrKs anaLysis in cancEr (DAMOKLE) on simulated and real data, showing that DAMOKLE does find subnetworks with significant differences in mutation frequency and that it provides novel insights into the molecular mechanisms of the disease not revealed by standard methods
In all cases we considered only the subnetwork with the highest differential coverage among the ones returned by DAMOKLE

Summary

Introduction

The analysis of molecular measurements from large collections of cancer samples has revolutionized our understanding of the processes leading to a tumour through somatic mutations, changes of the DNA appearing during the lifetime of an individual [1]. The availability of molecular measurements in a large number of samples for different cancer types have allowed comparative analyses of mutations in cancer [5, 10, 11]. Such analyses usually analyze large cohorts of different cancer types as a whole employing methods to find genes or subnetworks mutated in a significant fraction of tumours in one cohort, and analyze each cancer type individually, with the goal to identify: Hajkarim et al Algorithms Mol Biol (2019) 14:10 be assessed as significant only by the joint analysis of both sets of samples

Methods

Results

Conclusion