3D chromatin loops measured with Hi-C bring together SNP-SNP pairs engaging in epistatic interactions in GWAS data

Abstract

Motivation: Genome-wide association studies (GWAS) aim to uncover the genetic basis of traits and common diseases. Due to the large number of common variants, most studies use a single-locus approach. However, those fail to explain most of the heritability, especially for complex diseases. Epistatic interactions, where two or more loci have a synergistic effect on the phenotype, are worth investigating to improve our understanding of the genetic architecture of human disease. Most studies that have investigated epistatic association in GWAS focus on gene-gene interactions only. However, regulatory elements such as enhancers have the power of increasing and decreasing the expression level of their target genes, playing a fundamental role in determining their effects, also in relation to diseases. Thus, enhancer-promoter, or enhancer-enhancer interactions should be included in the search for GWAS epistasis. More and more studies show that the control of gene expression can occur over large genomic distances. Enhancers loop over to get in physical contact with their target genes. In vivo, enhancers and promoters are therefore found in close 3D spatial proximity. Chromatin loops can be detected using the chromosome conformation capture (3C) technique and its derivatives. Particularly, Hi-C combines 3C with next generation sequencing (NGS), identifying all contacts between all pairs of genomic regions. Results: In this study, we investigate GWAS epistatic effects of single nucleotide polymorphisms pairs (SNPs) engaging in long-range chromatin interactions. To this end, we overlay GWAS hits, using the T2D (type 2 diabetes) dataset from the WTCCC (Wellcome Trust Case-Control Consortium), with high resolution Hi-C maps. We show that chromatin loops are enriched for common variants, particularly when highly associated with the phenotype. Moreover, looping regions are associated with enhancer activity. We find three sets of SNP pairs engaging in epistatic interactions, on chromosomes 2, 3 and 12. The SNPs are found in either regions with high enhancer activity or in genes involved in metabolic pathways, which supports their potential role in type 2 diabetes (T2D).