Genome-wide Expression Quantitative Trait Loci: Results from the NHLBIs SABRe CVD Initiative. R. Joehanes1,2, T. Huan1, C. Yao1, X. Zhang1, S. Ying2, M. Feolo3, N. Sharopova3, T. Przytycka3, A. Sturcke3, A. A. Schaffer3, N. Heard-Costa4, H. Chen6, P. Liu5, R. Wang5, K. A. Woodhouse5, N. Raghavachari5, J. Dupuis4,6, A. D. Johnson1, C. J. O'Donnell1,7, P. J. Munson2, D. Levy1 1) Division of Intramural Research, National Heart, Lung and Blood Institute; the NHLBIs Framingham Heart Study, National Institutes of Health, Bethesda, MD, USA; 2) Mathematical and Statistical Computing Laboratory, Center of Information Technology, National Institutes of Health, Bethesda, MD, USA; 3) National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, USA; 4) School of Public Health, Boston University, Boston, MA, USA; 5) DNA Sequencing and Genomics Core, National Institutes of Health, Bethesda, MD, USA; 6) Department of Biostatistics, Boston University School of Public Health; 7) Division of Cardiology, Massachusetts General Hospital, Boston, MA, USA.
Introduction: Expression quantitative trait loci (eQTL) analysis has helped validate genotype-trait associations identified through genome-wide association studies (GWAS). Several eQTL studies have been performed, but many are derived from relatively small sample sizes. Methods: In this study, we performed an eQTL analysis using genotype and gene transcriptomic data from 5,257 Framingham Heart Study participants, the largest eQTL study to date. In total, we analyzed 39,315,185 1000G-imputed single nucleotide polymorphisms (SNPs) in association with whole blood-derived expression levels of 283,805 exons and 17,873 genes, measured in the Affymetrix Human Exon Array 1.0 ST platform. We used an additive regression model, adjusted for sex, age, family structure, complete blood counts, and technical covariates, such as batch. Cis eQTL is defined as eQTL 1 MB up and downstream of the corresponding transcript location. Results: At a false discovery rate <0.05, at the exon level, we identified 125,219,884 cis-eQTL and 64,066,764 trans-eQTL-exon pairs (21,694,034 unique SNPs and 282,730 unique exons). At the gene level, we detected 3,423,257 cis-eQTL and 36,218,111 trans-eQTL-gene pairs (9,435,973 unique SNPs and 17,873 unique genes). Of these, 3,447 exon-level and 2,172 gene-level eQTLs are reported as trait-associated SNPs in the NHGRI GWAS catalog, giving functional support to the GWAS results of various phenotypes. Among such phenotypes, platelet count and mean platelet volume are associated with several of the strongest eQTLs, such as rs1354034, rs16971217, rs10512472, rs12485738, rs505404, rs11602954, and rs17655730. Of these, rs1354034 in ARHGEF3 is a strong trans-eQTL that is associated with expression of at least 55 other transcripts across the genome. This finding lends support to the hypothesis that ARHGEF3 mRNA interactions are important in thrombopoiesis. In addition, we found strong eQTLs for 567 other phenotypes. For example, rs2247056 is a strong eQTL for triglycerides, rs3131379 for systemic lupus erythematosus, rs3916765 for type 2 diabetes, and rs3117582 for lung cancer. Our results will be available in NCBI GTEx database. Conclusion: Our eQTL results can be used to suggest functional elements relevant to numerous phenotypes. Our results also suggest that many of the SNPs identified by GWAS studies exert their influence to their respective phenotypes through mRNA expression levels of genes that may be distant from the SNP locations.
You may contact the first author (during and after the meeting) at