High-Resolution View of Genetic Architecture Underlying Autism
Objectives: A central goal in autism research is to identify the core set of highly penetrant, causative ASD genes among the growing list of genes associated with the disorder. Towards this, we have conducted systematic assessment of a large set of genes curated in the autism genetic database (AutDB) to generate a ranked list of candidate genes.
Methods: Using a revised gene-scoring algorithm that relies on evaluation of individual variants (Larsen et al., 2016), we have analyzed a total of 11,287 variants in 921 genes associated with ASD that were annotated from 1383 research articles (AutDB data freeze of June 2017). Our assessment criteria included significance of genetic association, family structure (simplex, multiplex, multi-generational, or consanguineous), zygosity (heterozygous, homozygous, or hemizygous), inheritance pattern (de novo or transmitted), the type of variant (missense, nonsense, etc.), and the functional effect of the variant.
Here, we present the evidence score and categorization of all ASD-linked genes catalogued in AutDB. The level of evidence in terms of number and type of genetic variants are broadly distributed in this dataset delineating distinct categories of ASD genes. Importantly, we report the identification of a set of 21 ASD genes that occupies more than two standard deviations (SDs) above the mean score of genes analyzed in this study. Using integrated bioinformatics analyses we further characterize the top ranking ASD gene set for their convergent biological function. We show that mouse prenatal and postnatal lethality genes were found to be enriched in this top-ranking ASD gene set (hypergeometric distribution test; p=3.0 x 10-5 and p=2.0 x 10-8 respectively) replicating an earlier finding of involvement of essential genes in ASD.
Conclusions: The evidence-based classification of ASD genes presented here is anticipated to help interpret newly identified genes and variations in ASD individuals.