Integrating epigenetic, genetic, and phenotypic data to uncover gene-region associations with triglycerides in the GOLDN study.

Background: There has been significant interest in investigating genome-wide and epigenome-wide associations with lipids. Testing at the gene or region level may improve power in such studies.

Methods: We analyze chromosome 11 cytosine-phosphate-guanine (CpG) methylation levels and single-nucleotide polymorphism (SNP) genotypes from the original Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study, aiming to explore the association between triglyceride levels and genetic/epigenetic factors. We apply region-based tests of association to methylation and genotype data, in turn, which seek to increase power by reducing the dimension of the gene-region variables. We also investigate whether integrating 2 omics data sets (methylation and genotype) into the triglyceride association analysis helps or hinders detection of candidate gene regions.

Results: Gene-region testing identified 1 CpG region that had been previously reported in the GOLDN study data and another 2 gene regions that are also associated with triglyceride levels. Testing on the combined genetic and epigenetic data detected the same genes as using epigenetic or genetic data alone.

Conclusions: Region-based testing can uncover additional association signals beyond those detected using single-variant testing.