Project summary Exons in the genome that lead to the introduction of premature truncation (stop) codon and mark these transcripts as targets for nonsense-mediated decay are called poison exons (PEs). These PEs are alternatively spliced throughout mouse and human neurodevelopment and function. Moreover, genetic variants that perturb the splicing of PEs have been associated with neurodevelopmental disorders (NDDs), including epilepsy, malformations of cortical development and a pediatric neurodegenerative disorder. These exons are also targeted by antisense oligonucleotides (ASO) based therapies in active clinical trials in rare genetic epilepsies. However, the prevailing contribution of these elements to neurological disease is unknown for several reasons, (1) we do not know how these genetic variants lead to ectopic PE splicing (2) the full repertoire of PEs used throughout neurodevelopment is unknown and (3) the genomic features that define these elements are not known. To address these challenges, we propose a multidisciplinary effort that integrates conventional splice reporters, proteomics, patient-specific stem cell models, combinatorial long-read single-cell sequencing and machine learning methods that will elucidate the characteristics of PEs used in neurodevelopment and the properties of genetic variants that lead to aberrant splicing and neurological disease. In Aim 1 of this study, we will determine how the PE spliceosome is perturbed in individuals with epilepsy who harbor pathogenic variants that cause ectopic SCN1A PE splicing. We will determine the RNA-binding proteins (RBPs) that lead to ectopic splicing and use ASOs to modulate splicing and correct the aberrant PE inclusion in patient-specific induced pluripotent stem cell (iPSC)-derived neurons. For novel PE discovery we have previously successfully used long-read transcriptome sequencing to detect these exons in iPSC-derived organoids and human brain tissue, and enhanced their discovery by inhibiting NMD with cycloheximide (CHX). In Aim 2, we will use MAS- ISO-seq which integrates both long-read and single-cell sequencing to generate cell-type specific maps of PE usage across neurodevelopment. Moreover, we will use high-depth short read-sequencing, including of CHX treated cells to boost the discovery of PEs. In Aim 3 we will determine the genomic features that define PEs and use this information to detect genetic variants that can perturb splicing of these exons. We will mine publicly available genome sequencing data from individuals with NDDs to identify additional individuals with putative pathogenic variants that alter PE splicing and develop a webtool for dissemination to the community. Collectively, the outcomes of this study will provide a comprehensive map of PEs used throughout neurodevelopment facilitating future studies in brain development and function, resolving undiagnosed NDDs and other neurological disorders, as well as therapeutic ASO targets for neurological di...