Abstract
Autism spectrum disorder (ASD) is a highly heterogeneous neurodevelopmental disorder characterized by impaired social communication coupled with stereotyped behaviors and restricted interests. Despite the high concordance rate for diagnosis, there is little information on the magnitude of genetic contributions to specific ASD behaviors. Using behavioral/trait severity scores from the Autism Diagnostic Interview-Revised (ADI-R) diagnostic instrument, we compared the phenotypic profiles of mono- and dizygotic twins where both co-twins were diagnosed with ASD or only one twin had a diagnosis. The trait distribution profiles across the respective twin populations were first used for quantitative trait association analyses using publicly available genome-wide genotyping data. Trait-associated single nucleotide polymorphisms (SNPs) were then used for case-control association analyses, in which cases were defined as individuals in the lowest (Q1) and highest (Q4) quartiles of the severity distribution curves for each trait. While all of the ASD-diagnosed twins exhibited similar trait severity profiles, the non-autistic dizygotic twins exhibited significantly lower ADI-R item scores than the non-autistic monozygotic twins. Case-control association analyses of twins stratified by trait severity revealed statistically significant SNPs with odds ratios that clearly distinguished individuals in Q4 from those in Q1. While the level of shared genomic variation is a strong determinant of the severity of autistic traits in the discordant non-autistic twins, the similarity of trait profiles in the concordantly autistic dizygotic twins also suggests a role for environmental influences. Stratification of cases by trait severity resulted in the identification of statistically significant SNPs located near genes over-represented within autism gene datasets.