Joint analysis of <i>de novo</i> mutations from autism spectrum disorder, schizophrenia, congenital heart disease, and other developmental disorders improves detection power and implicates shared molecular pathways and CNS processes

Rare exonic variant studies have previously implicated overlapping risk genes and pathways for autism spectrum disorder (ASD), severe, undiagnosed developmental disorders (UDDs), intellectual disability (ID), congenital heart disease (CHD), and schizophrenia (SCZ). Here, we use a two-trait Bayesian integrative analysis approach on 43 287 ASD, UDD/ID, CHD, and SCZ case trios to increase statistical power for gene discovery and to identify shared risk genes. At a posterior probability > 0.80, we identified 180 candidate risk genes for ASD, 315 for UDD/ID, 49 for CHD, and 47 for SCZ, including genes not previously reported, and also detected shared risk genes in pair-wise analyses. Gene set enrichment analysis of the ASD-UDD/ID, ASD-SCZ, and UDD/ID-SCZ shared risk genes overwhelmingly implicated gene sets associated with the synapse and epigenetic modification, while CHD-ASD shared risk genes were enriched in cell cycle phase transition gene sets, and CHD-UDD/ID shared risk genes implicated cardiac development. ASD-UDD/ID risk genes had elevated expression in interneurons and pyramidal cells, while ASD-UDD/ID and CHD-UDD/ID shared risk genes showed elevated connectivity in protein-protein interaction networks. Leveraging information across disorders with genetic overlap, both to increase power for candidate risk gene discovery and also as a method to elucidate shared genetic mechanisms.