Identification and Functional Assessment of Candidate Causal Cis -Regulatory Variants Underlying Electrocardiographic QT Interval GWAS Loci

BACKGROUND: Identifying causal variants among tens or hundreds of associated variants at each locus in genome-wide association studies is challenging. As the vast majority of genome-wide association studies variants are noncoding, sequence variation at cis -regulatory elements (CREs) affecting transcriptional expression of specific genes is a widely accepted molecular hypothesis. Following this hypothesis, combined with the observation that open chromatin is a universal hallmark of all types of CREs, we aimed to identify candidate causal cis -regulatory variants underlying QT interval genome-wide association studies loci. METHODS: Common variants in high linkage disequilibrium with genome-wide significant variants were identified using variant call format tools. Genome-wide maps of cardiac putative CREs were generated by MACS2-based peak calling in human cardiac left ventricular DNase I sequencing and Assay for Transposase-Accessible Chromatin using sequencing data sets ( n =13). Variant-CRE overlap was performed using custom tracks in the Table Browser tool at the UCSC Genome Browser. Luciferase reporter-based enhancer assays for variant-centered test elements were performed in mouse HL1 cardiomyocyte cells. Reporter activities of allelic pairs were compared using the Wilcoxon rank-sum test. RESULTS: At a dozen genome-wide association studies loci, selected for higher effect sizes and better understanding of the likely causal genes, we identified all genome-wide significant variants ( n =1401) and included all common variants (minor allele frequency >1%) in high linkage disequilibrium ( r 2 >0.9) with them as candidate variants ( n =3482). Candidate variants were filtered for overlap with cardiac left ventricular putative CREs to identify candidate causal cis -regulatory variants ( n =476), which were further assessed for being a known cardiac expression quantitative trait locus variant as additional functional evidence ( n =243). Functional evaluation of a subset of seven candidate variants by luciferase reporter-based enhancer assays in HL1 cells using variant-centered test elements led to the identification of 6 enhancer variants with significant allelic differences. CONCLUSIONS: These efforts have generated a comprehensive set of candidate causal variants expected to be enriched for cis -regulatory potential and thereby, explaining the observed genetic associations.