Federated analysis of the contribution of recessive coding variants to 29,745 developmental disorder patients from diverse populations

Autosomal recessive (AR) coding variants are a well-known cause of rare disorders. We quantified the contribution of these variants to developmental disorders (DDs) in the largest and most ancestrally diverse sample to date, comprising 29,745 trios from the Deciphering Developmental Disorders (DDD) study and the genetic diagnostics company GeneDx, of whom 20.4% have genetically-inferred non-European ancestries. The estimated fraction of patients attributable to exome-wide AR coding variants ranged from [~]2% to [~]18% across genetically-inferred ancestry groups, and was significantly correlated with the average autozygosity (r=0.99, p=5x10-6). Established AR DD-associated (ARDD) genes explained 90% of the total AR coding burden, and this was not significantly different between probands with genetically-inferred European versus non-European ancestries. Approximately half the burden in these established genes was explained by variants not already reported as pathogenic in ClinVar. We estimated that [~]1% of undiagnosed patients in both cohorts were attributable to damaging biallelic genotypes involving missense variants in established ARDD genes, highlighting the challenge in interpreting these. By testing for gene-specific enrichment of damaging biallelic genotypes, we identified two novel ARDD genes passing Bonferroni correction, KBTBD2 (p=1x10-7) and CRELD1 (p=9x10-8). Several other novel or recently-reported candidate genes were identified at a more lenient 5% false-discovery rate, including ZDHHC16 and HECTD4. This study expands our understanding of the genetic architecture of DDs across diverse genetically-inferred ancestry groups and suggests that improving strategies for interpreting missense variants in known ARDD genes may allow us to diagnose more patients than discovering the remaining genes.