RESUMEN
Predicting genotype-to-phenotype correlations from genomic variants has been challenging, particularly for genes that have a complex balance of dominant and recessive inheritance for phenotypes. Variants in NMDA receptor components GRIN1, GRIN2A, and GRIN2B cause a myriad of dominant disease phenotypes, with the most common being epilepsy and autism spectrum disorder. Starting from the analysis of a variant of uncertain significance (VUS, GRIN2A G760S), we realized the need for tools to map dominant variants for the components of the NMDA receptor. Some variants within GRIN1, GRIN2A, and GRIN2B exert dominant epilepsy and developmental delay, yet other amino acid variants are conserved and predicted to alter protein function but do not have dominant phenotypes. Common variant annotation tools are not powered to determine pathogenic dominant outcomes. To address this gap, we integrated sequence and structural analyses for GRIN1, GRIN2A, and GRIN2B. Using this approach, we determined that paralog homology mapping and topology can segregate dominant variants, with an elevation of intermolecular contacts between the subunits. Furthermore, demonstrating the general utility of our methodology, we show that 25 VUS within ClinVar also reach a dominant variant annotation, including the GRIN2A G760S variant. Our work suggests paralog homology and protein topology as a powerful strategy within the receptor complex to resolve dominant genetic variants relative to variants that would fit a recessive inheritance, requiring two damaging variants. These strategies should be tested in additional dominant genetic disorders to determine the broader utility.