Introduction: Alcohol intake modifies plasma lipid levels and such effects may be modulated by genetic variants. We use emerging statistical methods that extend well-established common variant approaches to characterize the role of aggregated rare and low-frequency variants in gene by alcohol consumption interactions associated with fasting plasma lipid levels.
Methods: Up to 247,870 exonic variants on the Illumina HumanExome BeadChip and fasting plasma triglycerides (TG), and high- and low-density lipoprotein cholesterol (HDL-c and LDL-c) were measured in 46,443 European Americans from 4 studies (the Atherosclerosis Risk in Communities (ARIC) study, the Framingham Heart Study, the Netherlands Epidemiology of Obesity Study and the Women’s Genome Health Study) of the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium.
Using the gene-based rareGE method, we conducted exome-wide gene-environment (GxE) tests with genetic main effects estimated as fixed and random effects, and a joint analysis of genetic main and GxE interaction effects. Rare and low-frequency (minor allele frequency ≤ 5%) functional variants, (i.e. frameshift, nonsynonymous, stop/gain, stop/loss, and splicing) were aggregated by genes. Two dichotomous self-reported alcohol consumption variables, current drinker (at least 1 drink per week, yes/no) and regular drinker (at least 2 drinks per week, yes/no) were considered. A sample size weighted Z-test (weighted Stouffer’s method) was used to meta-analyze study-specific p-values. Exome-wide significance level was set at p < 3.7*10-6 (0.05/13368 genes), using a Bonferroni procedure to correct for multiple testing.
Results: We identified 24 gene-lipid associations at 13 known lipid loci (within 500kb) harboring rare and low-frequency variants through the joint analysis. In ARIC, numerous genes (PCSK9, LPL, LIPG, ANGPTL4, APOB, APOC3-A5) remained significant after conditioning on common index single nucleotide polymorphisms (SNPs), suggesting an independent role for rare variants at loci highlighted by previous genome-wide association studies. However, no significant gene-alcohol interactions were observed with rare and low-frequency variants on TG, HDL-c or LDL-c.
Conclusion: This study applied new statistical approaches to investigate the role of rare and low-frequency variants in gene-alcohol consumption interactions on lipid levels. Results show promise for other larger scale studies analyzing rare variant GxE interactions.