Autism spectrum disorders (ASD) are believed to have genetic and environmental origins, yet in only a modest fraction of individuals can specific causes be identified. To identify further genetic risk factors, the study sequenced the exomes of 175 ASD cases and their parents, finding that fewer than half of the cases (46.3%) carry a missense or nonsense de novo variant, with the overall mutation rate only modestly higher than expected. However, there is significantly enriched connectivity among the proteins encoded by genes harboring de novo missense or nonsense mutations, and excess connectivity to prior ASD genes of major effect, suggesting a subset of observed events are relevant to ASD risk. The small increase in de novo mutation rate, combined with the connections among the proteins and to ASD, supports a limited but important role for de novo point mutations, similar to that of de novo copy number variants. Genetic models suggest that most de novo events are not connected to ASD, while those that confer risk are distributed across many genes and are incompletely penetrant. The study supports polygenic models where spontaneous coding mutations in many genes increase ASD risk by 5 to 20-fold. Results from de novo events and a large case-control study provide strong evidence that CHD8 and KATNAL2 are genuine autism risk factors. The study found that de novo mutations are not significantly elevated in ASD, but there is a significant enrichment of connectivity among proteins encoded by genes with de novo mutations and to prior ASD genes. The study also found that paternal and maternal age are strong predictors of the number of de novo events per offspring. The study used exome sequencing and pathway analyses to evaluate the functional impact of de novo mutations, finding that some genes with multiple de novo mutations may be candidates for ASD risk. However, simulations suggest that two de novo hits are not sufficient to define a gene as a conclusive risk factor. The study also found that de novo mutations are not significantly associated with verbal, nonverbal, or full-scale IQ. The study concludes that de novo mutations may contribute to ASD risk, but their role is limited. The study also supports the idea that de novo SNVs may combine with other risk factors rather than fully cause disease. The study found that CHD8 and KATNAL2 are likely genuine autism susceptibility genes. The study highlights the challenge of establishing individual genes as conclusive risk factors for ASD, which will require larger sample sizes and deeper integration with inherited variation.Autism spectrum disorders (ASD) are believed to have genetic and environmental origins, yet in only a modest fraction of individuals can specific causes be identified. To identify further genetic risk factors, the study sequenced the exomes of 175 ASD cases and their parents, finding that fewer than half of the cases (46.3%) carry a missense or nonsense de novo variant, with the overall mutation rate only modestly higher than expected. However, there is significantly enriched connectivity among the proteins encoded by genes harboring de novo missense or nonsense mutations, and excess connectivity to prior ASD genes of major effect, suggesting a subset of observed events are relevant to ASD risk. The small increase in de novo mutation rate, combined with the connections among the proteins and to ASD, supports a limited but important role for de novo point mutations, similar to that of de novo copy number variants. Genetic models suggest that most de novo events are not connected to ASD, while those that confer risk are distributed across many genes and are incompletely penetrant. The study supports polygenic models where spontaneous coding mutations in many genes increase ASD risk by 5 to 20-fold. Results from de novo events and a large case-control study provide strong evidence that CHD8 and KATNAL2 are genuine autism risk factors. The study found that de novo mutations are not significantly elevated in ASD, but there is a significant enrichment of connectivity among proteins encoded by genes with de novo mutations and to prior ASD genes. The study also found that paternal and maternal age are strong predictors of the number of de novo events per offspring. The study used exome sequencing and pathway analyses to evaluate the functional impact of de novo mutations, finding that some genes with multiple de novo mutations may be candidates for ASD risk. However, simulations suggest that two de novo hits are not sufficient to define a gene as a conclusive risk factor. The study also found that de novo mutations are not significantly associated with verbal, nonverbal, or full-scale IQ. The study concludes that de novo mutations may contribute to ASD risk, but their role is limited. The study also supports the idea that de novo SNVs may combine with other risk factors rather than fully cause disease. The study found that CHD8 and KATNAL2 are likely genuine autism susceptibility genes. The study highlights the challenge of establishing individual genes as conclusive risk factors for ASD, which will require larger sample sizes and deeper integration with inherited variation.