Fast search of third-order epistatic interactions on CPU and GPU clusters

Abstract

Genome-Wide Association Studies (GWASs), analyses that try to find a link between a given phenotype (such as a disease) and genetic markers, have been growing in popularity in the recent years. Relations between phenotypes and genotypes are not easy to identify, as most of the phenotypes are a product of the interaction between multiple genes, a phenomenon known as epistasis. Many authors have resorted to different approaches and hardware architectures in order to mitigate the exponential time complexity of the problem. However, these studies make some compromises in order to keep a reasonable execution time, such as limiting the number of genetic markers involved in the interaction, or discarding some of these markers in an initial filtering stage. This work presents MPI3SNP, a tool that implements a three-way exhaustive search for cluster architectures with the aim of mitigating the exponential growth of the run-time. Modern cluster solutions usually incorporate GPUs. Thus, MPI3SNP includes implementations for both multi-CPU and multi-GPU clusters. To contextualize the performance achieved, MPI3SNP is able to analyze an input of 6300 genetic markers and 3200 samples in less than 6 min using 768 CPU cores or 4 min using 8 NVIDIA K80 GPUs. The source code is available at https://github.com/chponte/mpi3snp .

MIDESP: Mutual Information-Based Detection of Epistatic SNP Pairs for Qualitative and Quantitative Phenotypes

“…In this paper, we introduce a novel information-theory-based method, MIDESP, for the detection of epistatic interactions using genotype–phenotype datasets. MIDESP is able to analyze both qualitative as well as quantitative phenotypes, unlike previous information theoretical methods [ 39 , 41 , 42 , 43 , 44 , 45 , 46 ], which are only applicable to datasets with qualitative phenotypes. Furthermore, our method takes into account the effect of dataset-dependent background associations and eliminates them to some extent using the average product correction (APC) technique [ 53 ] to separate correct/functional epistatic signals from those of false positives.…”

“…Today, it is well established that mutual information is an appropriate metric to measure the association between SNPs and qualitative (case–control) phenotypes [ 39 , 44 , 46 , 74 , 75 , 76 , 77 ]. However, we apply here for the first time this metric to quantitative traits.…”

“…In previous studies [ 39 , 42 , 46 ], the epistatic interaction between an SNP pair, and , and a qualitative phenotype, Y , has been successfully identified by employing the MI metric for which Equation ( 1 ) is extended based on the joint entropy as: where is the joint entropy of the SNPs and as well as the phenotype Y . However, the concept of MI has not yet been applied to measure the epistatic interaction between an SNP pair and a quantitative phenotype.…”

“…Several of these methods use information-theory-based measures such as mutual information to quantify epistatic interactions [ 39 , 41 , 42 , 43 , 44 , 45 , 46 ]. These measures consider the SNPs and phenotypes as random variables, which allows them to quantify the amount of information, or uncertainty, that is inherent to an SNP or a phenotype and to compute how much information is shared between them, and thereby the strength of association [ 42 ].…”

Fast and Accurate Exhaustive Higher-Order Epistasis Search with BitEpi

“…Capable of processing higher-order interactions more efficiently, MDR [20] (Multi-factor Dimensionality Reduction) is an extensive epistasis analysis platform offering parallel exhaustive search functionality. Improving on the algorithmic implementation further, MPI3SNP [21] adapts the bitwise approach used by BOOST. However, with MPI3SNP being limited to 3-SNV searches, the need for a fast higher-order search remains unaddressed.…”