Molecular diagnoses and candidate gene identification in the congenital heart disease cohorts of the 100,000 genomes project

Congenital heart disease (CHD) describes a structural cardiac defect present from birth. A cohort of participants recruited to the 100,000 Genomes Project (100 kGP) with syndromic CHD (286 probands) and familial CHD (262 probands) were identified. "Tiering" following genome sequencing data analysis prioritised variants in gene panels linked to participant phenotype. To improve diagnostic rates in the CHD cohorts, we implemented an agnostic de novo Gene Discovery Pipeline (GDP). We assessed de novo variants (DNV) for unsolved CHD participants following filtering to select variants of interest in OMIM-morbid genes, as well as novel candidate genes. The 100kGP CHD cohorts had low rates of pathogenic diagnoses reported (combined CHD "solved" 5.11% (n = 28/548)). Our GDP provided diagnostic uplift of nearly one third (1.28% uplift; 5.11% vs. 6.39%), with a new or potential diagnosis for 9 additional participants with CHD. When a filtered DNV occurred within a non-morbid gene, our GDP prioritised biologically-plausible candidate CHD genes (n = 79). Candidate variants occurred in both genes linked to cardiac development (e.g. AKAP13 and BCAR1) and those currently without a known role (e.g. TFAP2C and SETDB1). Sanger sequencing of a cohort of patients with CHD did not identify a second de novo variant in the candidate dataset. However, literature review identified rare variants in HMCN1, previously reported as causative for pulmonary atresia, confirming the approach utility. As well as diagnostic uplift for unsolved participants of the 100 kGP, our GDP created a dataset of candidate CHD genes, which forms an important resource for further evaluation.