Age-aware genotype-phenotype architecture across 87 genetic neurodevelopmental disorders

Neurodevelopmental disorders (NDDs) are clinically heterogeneous, multisystem disorders, making prognosis challenging, even after a molecular diagnosis is obtained. In this study, we assembled a harmonized, age-aware genotype-phenotype atlas using data from the Simons Searchlight cohort. We analyzed 1,210 probands with pathogenic variants across 87 NDDs (75 gene, 12 recurrent copy-number variant disorders). Mapping 168 phenotypes onto 14 top-level Human Phenotype Ontology (HPO) categories showed that virtually all NDDs exhibited broad, multisystem involvement. In contrast, higher-resolution phenotype analyses uncovered 49 significant gene-phenotype enrichments (Bonferroni P < 3.9x10-), for example, microcephaly in DYRK1A, macrocephaly in PPP2R5D, and distinct seizure-subtype patterns in STXBP1 and SLC6A1. We calculated 1,368 age-dependent penetrance trajectories across 36 disorders and 139 phenotypes and use Cox models to identify 43 NDD-specific onset pattern (e.g., six times earlier absence seizures in SYNGAP1 compared to other NDDs). Here we show that combining calibrated prevalence with developmental timing converts genetic findings into quantitative, gene-specific risk profiles. This atlas enables earlier recognition, gene-informed surveillance, and more rational trial design. The atlas is available for interactive exploration in the NDD-Portal (https://lalresearchgroup.org).