Genome-wide identification of novel genetic variants associated with erythrocyte membrane fatty acids. A. E. Locke1, A. U. Jackson1, A. Stancáková2, Y. Wu3, T. M. Teslovich1, C. Fuchsberger1, N. Narisu4, P. Chines4, R. Welch1, H. M. Stringham1, X. L. Sim1, J. Huyghe1, M. Civelek5, N. K. Saleem5, A. He6, C. Tilford6, P. Gargalovic6, T. Kirchgessner6, A. J. Lusis5, K. Mohlke3, M. Boehnke1, M. Laakso2 1) Department of Biostatistics and Center for Statistical Genetics, University of Michigan, Ann Arbor, MI; 2) Department of Medicine, Kuopio University Hospital and University of Eastern Finland, Kuopio, Finland; 3) Department of Genetics, University of North Carolina, Chapel Hill, NC; 4) Genome Technology Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD; 5) Department of Medicine, University of California, Los Angeles, CA; 6) Bristol-Myers Squibb, Pennington, NJ.

   Several genes have been associated with fatty acid levels. However, most studies have measured fatty acids from plasma, which have a short half-life (~2-4 minutes) and are heavily influenced by diet. In contrast, erythrocyte membrane bound fatty acids (EMFAs) persist for life of the erythrocyte (~ 3 months), are less influenced by dietary intake, and thus can better reflect genetic factors contributing to fatty acid synthesis and metabolism. As part of the METabolic Syndrome In Men (METSIM) study, we used gas chromatography to measure 20 fatty acids (six saturated, five mono-unsaturated, and nine poly-unsaturated), the cumulative measures of all saturated, mono-unsaturated, and poly-unsaturated fatty acids, and six ratios measuring enzymatic activity in 2,299 adult males from Kuopio, Finland. Following genotyping with Illumina Omni Express and Illumina Metabochip and imputation using the GoT2D reference panel, we performed single variant and gene-based association analysis with each of the 28 fatty acid measures. Twenty-four of the 28 traits were associated with at least one variant, and in total 45 single variant associations at 21 different loci reached genome-wide significance. In addition to established associations with fatty acid desaturases (FADS1-2-3), ELOVL2, and PDXDC1, we also identified associations near strong functional candidate genes: fatty acid elongases (ELOVL5 and ELOVL6), acyltransferases (LPCAT3, AGPAT4, AGPAT5, and CERS4), and palmitoyltransferases (ZDHHC21, CPT1A, and SPTLC3). In conditional analyses, we identified three genomic regions with multiple associations for six different traits. Finally, we included putative functional variants (protein truncating and missense) in gene-based tests and identified low frequency variants in ABHD3 significantly associated with myristic acid. We used eQTL results from subcutaneous adipose tissue expression data from 1,410 METSIM individuals to identify potential functional candidate genes at these loci. In reciprocal conditional analyses, the fatty acid GWAS SNP at four loci accounted for the expression signal of a candidate gene (FADS1, SCD, PDXDC1, CERS4) and vice versa. Measurement and association testing of fatty acids derived from erythrocyte membranes, along with expression data from the same individuals, has considerably enhanced understanding of the biological mechanisms of fatty acid metabolism.

You may contact the first author (during and after the meeting) at