Abstract
Previous genetic studies in Autism Spectrum Disorder (ASD) identified hundreds of high-confidence ASD genes enriched with likely deleterious protein-coding de novo mutations (DNMs). Multiple studies also demonstrated that DNMs in the non-coding genome can contribute to ASD risk. However, the identification of individual risk genes enriched with noncoding DNMs has remained largely unexplored. We analyzed two datasets with over 5000 ASD families to assess the contribution of noncoding DNMs. We used two methods to assess statistical significance for noncoding DNMs: a point-based test that analyzes sites that are likely functional, and a segment-based test that analyzes 1 kb genomic segments with segment-specific background mutation rates. We found that coding and noncoding DNMs in SCN2A are associated with ASD risk. Further application of these approaches on large-scale whole genome sequencing data will help identify additional candidate ASD risk genes.