Background Only a small fraction of the heritability of coronary artery disease (CAD) has been explained by common variants identified by genome-wide association studies. Among the stones to be turned in the hunt for the missing heritability of CAD are gene-gene interactions. We investigated whether interactions between common alleles in genes and pathways of known importance to cardiovascular regulation may contribute to the heritability of CAD.
Methods 2101 CAD cases and 2426 controls of Caucasian origin recruited into Wellcome Trust Case Control Consortium were genotyped using 50 K IBC gene-centric array containing 45 707 single nucleotide polymorphisms (SNPs) of the highest biological relevance to cardiovascular system. After applying appropriate quality control filters, 11 332 common (minor allele frequency >10%), independent (r2 linkage disequilibrium coefficient of ≤0.5) were included in pair-wise SNP-SNP interaction analysis using two complementary statistical approaches: logistic regression (PLINK and INTERSNP software packages) and Bayesian model (BEAM software).
Results None of the analysed SNP-SNP interactions was statistically significant after correction for multiple testing (p=7.8×10-10). The most significant interaction identified in this analysis was between rs727139 (KCNH8) on chromosome 3 and rs11167496 (PDGFRB) on chromosome 5 (p=2.45×10−8). Analysis of subsets of SNPs pre-selected based on their nominal association with CAD (p<0.05) or molecular functionality (non-synonymous SNPs) did not contribute more significant findings than investigation of random set of SNPs.
Conclusion Our analysis suggests that common SNP-SNP interactions are unlikely to account for a large proportion of the missing heritability of CAD.
- gene-centric array