Network graph analysis of gene-gene interactions in genome-wide association study data

Abstract

Most common complex traits, such as obesity, hypertension, diabetes, and cancers, are known to be associated with multiple genes, environmental factors, and their epistasis. Recently, the development of advanced genotyping technologies has allowed us to perform genome-wide association studies (GWASs). For detecting the effects of multiple genes on complex traits, many approaches have been proposed for GWASs. Multifactor dimensionality reduction (MDR) is one of the powerful and efficient methods for detecting high-order gene-gene (GxG) interactions. However, the biological interpretation of GxG interactions identified by MDR analysis is not easy. In order to aid the interpretation of MDR results, we propose a network graph analysis to elucidate the meaning of identified GxG interactions. The proposed network graph analysis consists of three steps. The first step is for performing GxG interaction analysis using MDR analysis. The second step is to draw the network graph using the MDR result. The third step is to provide biological evidence of the identified GxG interaction using external biological databases. The proposed method was applied to Korean Association Resource (KARE) data, containing 8838 individuals with 327,632 single-nucleotide polymorphisms, in order to perform GxG interaction analysis of body mass index (BMI). Our network graph analysis successfully showed that many identified GxG interactions have known biological evidence related to BMI. We expect that our network graph analysis will be helpful to interpret the biological meaning of GxG interactions.

Keywords: gene-gene interaction; generalized multifactor dimensionality reduction; genome-wide association study; graph analysis; graphic processing units; network graph.

Fig. 1

Visualized result of significant interactions that have their cross-validation consistency ≥9. Arranged for readability: gray background indicates hub node, and red, white, blue, and yellow names indicate that they are identified for their relation with obesity, single-nucleotide polymorphism, gene, and unidentified gene locus, respectively.

Fig. 2

A visualization of gene-gene interaction interpretation with biological knowledge. Two red circles denote two single-nucleotide polymorphisms (SNPs) within a two-way interaction, and purple circles denote corresponding genes against two SNPs. Gray circles denote diseases that are known to be related with both of the genes. Yellow and orange circles denote a disrupted transcriptional factor by both of the two SNPs and gene sets including genes from both SNPs, respectively.