Allele-specific NKX2-5 binding underlies multiple genetic associations with human electrocardiographic traits

We conducted a genome-wide analysis to identify regulatory variants affecting the binding of NKX2-5, a core cardiac development transcription factor, and investigated their role in cardiac gene expression and EKG phenotypes. We generated iPSC-derived cardiomyocytes (iPSC-CMs) from a pedigree of seven whole-genome sequenced individuals, and profiled them with a variety of functional genomic assays including RNA-Seq, ATAC-Seq, and ChIP-Seq of both histone modification H3K27ac and NKX2-5. After establishing that iPSC-CMs recapitulated cardiomyocyte-specific expression and epigenetic signatures, and that genetic variants affected the variability of molecular phenotypes across the iPSC-CM lines, we identified heterozygous sites that showed allele-specific effects (ASE). We then investigated NKX2-5 ASE variants in detail by examining whether they altered cardiac TF motifs, and whether they were enriched for eQTLs and EKG GWAS-SNPs. Our data reveal that variation affecting the binding of NKX2-5 and other cardiac TFs likely serves as a molecular mechanism underlying control of numerous EKG loci across the genome, and that fine-mapping approaches, combined with molecular phenotype data from iPSC-CMs, can be used to prioritize causal variants in EKG GWAS loci. We selected seven individuals of Asian and European descent in the iPSCORE resource that are part of a three-generational family. Our study design included three genetically unrelated subjects and two parent-offspring quartets, which enabled us to examine the inheritance of genetic effects.We generated and analyzed 56 RNA-Seq (iPSCs: 29 independent samples; iPSC-CMs: 26 independent samples and 1 technical replicate), 48 ChIP-Seq of histone modification H3K27ac (iPSCs: 17 samples and 4 technical replicates; iPSC-CMs: 25 samples and 2 technical replicates), 15 ChIP-seq of NKX2-5 (iPSC-CMs: 12 samples and 3 technical replicates and 37 ATAC-Seq (iPSCs: 12 samples and 5 technical replicates; iPSC-CMs: 11 samples and 9 technical replicates). Raw data requires controlled access and is deposited at dbGaP (phs000924 and phs001325). Data for RNA-Seq, H3K7ac ChIP-Seq and 21/37 ATAC-Seq samples were deposited under the accession GSE125540.