We sequenced exomes from more than 2,500 simplex families each having a child with an autistic spectrum disorder (ASD). By comparing affected to unaffected siblings, we estimate that 13% of de novo (DN) missense mutations and 42% of DN likely gene-disrupting (LGD) mutations contribute to 12% and 9% of diagnoses, respectively. Including copy number variants, coding DN mutations contribute to about 30% of all simplex and 45% of female diagnoses. Virtually all LGD mutations occur opposite wild-type alleles. LGD targets in affected females significantly overlap the targets in males of lower IQ, but neither overlaps significantly with targets in males of higher IQ. We estimate that LGD mutation in about 400 genes can contribute to the joint class of affected females and males of lower IQ, with an overlapping and similar number of genes vulnerable to causative missense mutation. LGD targets in the joint class overlap with published targets for intellectual disability and schizophrenia, and are enriched for chromatin modifiers, FMRP-associated genes and embryonically expressed genes. Virtually all significance for the latter comes from affected females.

Autism spectrum disorder (ASD) is a heterogeneous disease where efforts to define subtypes behaviorally have met with limited success. Hypothesizing that genetically based subtype identification may prove more productive, we resequenced the ASD-associated gene CHD8 in 3,730 children with developmental delay or ASD. We identified a total of 15 independent mutations; no truncating events were identified in 8,792 controls, including 2,289 unaffected siblings. In addition to a high likelihood of an ASD diagnosis among patients bearing CHD8 mutations, characteristics enriched in this group included macrocephaly, distinct faces, and gastrointestinal complaints. chd8 disruption in zebrafish recapitulates features of the human phenotype, including increased head size as a result of expansion of the forebrain/midbrain and impairment of gastrointestinal motility due to a reduction in post-mitotic enteric neurons. Our findings indicate that CHD8 disruptions define a distinct ASD subtype and reveal unexpected comorbidities between brain development and enteric innervation.

Copy number variants (CNVs) are associated with many neurocognitive disorders; however, these events are typically large and the underlying causative gene is unclear. We created an expanded CNV morbidity map from 29,085 children with developmental delay versus 19,584 healthy controls, identifying 70 significant CNVs. We resequenced 26 candidate genes in 4,716 additional cases with developmental delay or autism and 2,193 controls. An integrated analysis of CNV and single-nucleotide variant (SNV) data pinpointed ten genes enriched for putative loss of function. Patient follow-up on a subset identified new clinical subtypes of pediatric disease and the genes responsible for disease-associated CNVs. This includes haploinsufficiency of SETBP1 associated with intellectual disability and loss of expressive language and truncations of ZMYND11 in patients with autism, aggression and complex neuropsychiatric features. This combined CNV and SNV approach facilitates the rapid discovery of new syndromes and neuropsychiatric disease genes despite extensive genetic heterogeneity.

To assess the relative impact of inherited and de novo variants on autism risk, we generated a comprehensive set of exonic single nucleotide variants (SNVs) and copy number variants (CNVs) from 2,377 autism families. We find that private, inherited truncating SNVs in conserved genes are enriched in probands (odds ratio=1.14, p=0.0002) compared to unaffected siblings, an effect with significant maternal transmission bias to sons. We also observe a bias for inherited CNVs, specifically for small (<100 kbp), maternally inherited events (p=0.01) that are enriched in CHD8 target genes (p=7.4×10−3). Using a logistic regression model, we show that private truncating SNVs and rare, inherited CNVs are statistically independent autism risk factors, with odds ratios of 1.11 (p=0.0002) and 1.23 (p=0.01), respectively. This analysis identifies a second class of candidate genes (e.g., RIMS1, CUL7, and LZTR1) where transmitted mutations may create a sensitized background but are unlikely to be completely penetrant.