Assessing the contribution of rare variants to congenital heart disease through a large-scale case-control exome study

Abstract

Several studies have demonstrated the value of large-scale human exome and genome data analysis, to maximise gene discovery in rare diseases. Using this approach, we have analysed the exomes of 4,747 cases and 52,881 controls, to identify single genes and digenic interactions which confer a substantial risk of congenital heart disease (CHD). We identified both rare loss-of-function and missense coding variants in ten genes which reached genome-wide significance (Bonferroni adjustedP< 0.05) and an additional four genes with a significant association at a false discovery rate (FDR)threshold of 5%. We highlight distinct genetic contributions to syndromic and non-syndromic CHD at both single gene and digenic level, by independently analysing probands from these two groups. In addition, by integrative analysis of exome data with single-cell transcriptomics data from human embryonic hearts, we identified cardiac-specific cells as well as putative biological processes underlying the pathogenesis of CHD. In summary, our findings strengthen the association of known CHD genes, and have identified additional novel disease genes and digenic interactions contributing to the aetiology of CHD.

TAK1 operates at the primary cilium in non-canonical TGFB/BMP signaling to control heart development

“…To investigate the association between TAK1 , TAB2, PRKACA and CHD, we initially analyzed exome sequencing (ES) data from a cohort of 1,471 sCHD patients, 2,405 nsCHD patients and 45,082 controls 20 . We calculated the burden of rare (MAF<0.001) synonymous (SYN) and protein altering variants (PAVs) between the two patient cohorts and controls and analyzed the difference using Fisher’s exact test.…”

“…Variant burden was calculated using ES data from a case-control dataset of 4,747 CHD cases and 52,881 controls 20 . Briefly, case and control data were processed and harmonized using the same alignment (BWA v0.7), variant calling (GATK v4.1), variant annotation (VEP API 96) and quality control (Hail v0.2) pipelines.…”

“…Significant overlap was calculated using hypergeometric statistics (http://nemates.org/MA/progs/overlap_stats.html). A representation factor was calculated as the number of overlapping genes, divided by the expected number of overlapping genes drawn from two independent groups: RF=x/((n*D)/N), where x=number of overlapping genes, n=genes in group 1, D=genes in group 2, N=genes in genome (20,000).…”