Identification and characteristics of common genetic variants controlling transcript isoform variation in the Framingham Heart Study. X. Zhang1, R. Joehanes1,2, T. Huan1, P. Munson2, A. Johnson1, D. Levy1, C. ODonnell1,3 1) NHLBIs Framingham Heart Study, Framingham, MA; 2) Center for Information Technology, National Institutes of Health, Bethesda, MD; 3) Division of Cardiology, Massachusetts General Hospital, Boston, MA.
Introduction: Alternative splicing (AS) affects ~80% of human pre-mRNAs. Splicing quantitative trait loci (sQTLs) contribute to phenotypic differences among individuals and may have important roles in disease susceptibility. However, the available evidence from sQTL studies is derived from small sample sizes largely from Epstein-Barr virus transformed lymphoblastoid cell lines (LCLs) which may not represent in vivo environments. We performed genome-wide screening to identify SNPs controlling splicing in whole blood collected from the community-based Framingham Heart Study (FHS). Methods: 5,626 FHS participants were included. Total RNA was isolated from PAXgene whole blood samples. Expression levels of 17,873 genes and 283,805 exons were measured using Affymetrix Human Exon 1.0 ST arrays. Common SNPs (MAF 0.01) within a 50kb region flanking either side of the gene were selected from a 1000 genomes imputed SNP dataset. A cis-sQTL analysis for each of the genes and exons was conducted using an additive regression model adjusted for age, sex and family structure. Bonferroni-correction was used to account for multiple testing. Results: On average there were 553 SNPs located within 50kb of each gene, and 16 exons per gene. 6,137 genes were significantly associated with at least one cis SNP (P < 5.2e-09) at the whole gene level. For 2,865 genes without association with any SNP at the gene level, at least one of the exons harbor one cis-sQTL at P < 2.8e-10. There were a total of 672,846 cis-sQTLs, suggesting that a large proportion of genes are affected by splicing regulatory variants. ~70% of these sQTLs are in intergenic or intronic and 5% are near a 5 promoter region. We identified 32 sQTLs in a 5' splice donor or 3' splice acceptor site. When we examined the NHGRI GWAS catalog, 602 unique sQTLs are found to be associated with 263 disease traits, indicating that many disease variants might affect pre-mRNA splicing. Furthermore, 40% of these 2,865 genes with specific cis-sQTL associations overlap with known AS events. These 2,865 genes are enriched for RNA-binding and alternative splicing GO terms. Conclusion: Many strong common cis-acting regulatory variants affect the splicing patterns of genes in a large population. Many genes show significant genetically controlled differences in splice-site usage. Our study provides a splicing-QTL catalog for researchers to discover functional sQTLs that control transcript isoforms implicated in common diseases.
You may contact the first author (during and after the meeting) at