Gene by gene interaction analyses identifies evidence of epistasis and new candidate genes that influence intraocular pressure in the general population

Abstract

Purpose: Genome-wide association studies (GWAS) have identified several loci determining intraocular pressure (IOP), many of which were also associated with susceptibility to primary open angle glaucoma. However, these variants collectively only explain a small proportion of the expected genetic effects for IOP. Epistatic interaction between genetic loci is one of the proposed hypotheses to explain missing heritability. The purpose of this study is to assess the importance of gene-gene epistatic interactions relating to IOP in population-based studies.

Methods: GWAS data from 15 cohorts involving about 28,000 subjects from the International Glaucoma Genetic Consortium (IGGC) were included in this study. A total of 12 single nucleotide polymorphism (SNPs) from the seven loci most strongly associated with IOP identified in the previous GWAS were selected a priori. Pairwise epistatic interactions between these 12 SNPs and all other SNPs in the HapMap2 dataset were assessed by calculating the statistical interaction term using linear regression models for each cohort independently. These terms were then combined by means of a fixed-effect inverse variance meta-analysis. The threshold of significance was set ar 4.1 x 10^-9 .

Results: Several associations met the adjusted genome-wide significance threshold (alpha=4.1E-09). The strongest associations were observed between rs10258482 (within the CAV1 gene) and SNPs within or in immediate proximity to the TOX2 (p=2.5E-11), TBC1D1 (p=2.2E-10), TNKS (2.8E-10) and NRXN3 (p=7.8E-10) genes. Previous works suggest that the last three genes are involved in fat metabolism and were associated with body mass and obesity. SNPs located on chromosome 4 showed moderate eQTL effects over the TBC1D1 (p=0.007) gene, but also had eQTL effects of similar magnitude over other adjacent genes.

Conclusions: This work has identified statistically significant associations of synergistic interaction between SNPs and IOP and illustrates the importance of using other approaches than study of linear associations of common variants for traits of interest, to include other potential sources of heritability. These results suggest that multi-locus effects are important and may need to be investigated more thoroughly in the future.