Abstract
Genome-wide association studies have identified numerous variants associated with lipid levels; yet, the majority are located in non-coding regions with unclear mechanisms. In the Insulin Resistance Atherosclerosis Family Study (IRASFS), heritability estimates suggest a strong genetic basis: low-density lipoprotein (LDL, h(2) = 0.50), high-density lipoprotein (HDL, h(2) = 0.57), total cholesterol (TC, h(2) = 0.53), and triglyceride (TG, h(2) = 0.42) levels. Exome sequencing of 1,205 Mexican Americans (90 pedigrees) from the IRASFS identified 548,889 variants and association and linkage analyses with lipid levels were performed. One genome-wide significant signal was detected in APOA5 with TG (rs651821, P(TG) = 3.67 × 10(-10), LOD(TG) = 2.36, MAF = 14.2%). In addition, two correlated SNPs (r(2) = 1.0) rs189547099 (P(TG) = 6.31 × 10(-08), LOD(TG) = 3.13, MAF = 0.50%) and chr4:157997598 (P(TG) = 6.31 × 10(-08), LOD(TG) = 3.13, MAF = 0.50%) reached exome-wide significance (P < 9.11 × 10(-08)). rs189547099 is an intronic SNP in FNIP2 and SNP chr4:157997598 is intronic in GLRB. Linkage analysis revealed 46 SNPs with a LOD > 3 with the strongest signal at rs1141070 (LOD(LDL) = 4.30, P(LDL) = 0.33, MAF = 21.6%) in DFFB. A total of 53 nominally associated variants (P < 5.00 × 10(-05), MAF ≥ 1.0%) were selected for replication in six Mexican-American cohorts (N = 3,280). The strongest signal observed was a synonymous variant (rs1160983, P(LDL) = 4.44 × 10(-17), MAF = 2.7%) in TOMM40. Beyond primary findings, previously reported lipid loci were fine-mapped using exome sequencing in IRASFS. These results support that exome sequencing complements and extends insights into the genetics of lipid levels.