De Novo Variation in Autism by Sex and Diagnostic Status in 41,367 Parent-Child Trios

Autism shows a consistent sex bias, yet how sex shapes <i>de novo</i> variant (DNV) risk across coding and noncoding sequence remains unclear. We analyzed DNVs in 41,367 parent-child sequenced trios from three autism family-based cohorts and compared DNV characteristics and enrichment patterns in males and females. Importantly, these trios consisted of some trios with individuals with autism and some without autism. We developed a new sex-aware DNV caller and performed intensive, feature-based investigation of each candidate DNV to produce a high-confidence callset. We identified enrichment of missense and loss-of-function (LOF) DNVs both overall and within known autism-related genes (i.e., SFARI genes). Gene-specific enrichment analyses revealed twelve genes that were exome-wide significant and specific to males, for significance, including <i>FOXP1, SMAD6, AUTS2, CCDC168, PIEZO1, EML6, ZNF84, IGSF23, OTOG, SLC6A1, GIGYF1</i>, and <i>FREM3</i> and four genes that were specific to females, for significance, including <i>TAOK1, MECP2, DDX3X</i>, and <i>TBL1XR1</i> within a variant class. Direct comparisons of DNVs in males and females revealed <i>GABBR2</i> as the only gene trending toward enrichment in the direct males with autism comparison to females with autism. Finally, we analyzed promoters and identified a single significant promoter region (p = 3.8×10^-13), associated with the <i>WDR74</i> gene, with the signal driven by DNVs observed in males with autism. Surprisingly, the noncoding RNA gene <i>RNU2-2</i> lies within this significant <i>WDR74</i> promoter and accounted for most of the DNVs in the region. <i>RNU2-2</i> DNVs were present in 0.2% of males with autism, and several are predicted to potentially alter RNA folding. We also observed <i>RNU2-2</i> DNVs in 0.2% of females with autism, including two DNVs that were recurrent (i.e., shared) with unrelated, affected males. Notably, <i>RNU2-2</i> DNVs were detected in 0.1% of unaffected males and were not observed in unaffected females. Together, these results suggest that although <i>RNU2-2</i> does not show a sex bias, it contributes to autism risk, which is intriguing due to a prior study implicating <i>RNU2-2</i> in a severe neurodevelopmental disorder.