Analysis of the Autism Phenotype: Identification of Distinct Sub-populations

Background: Autism is a neurodevelopmental disorder affecting the three core areas of communication, social interaction and restrictive repetitive behaviors.  It affects 1 in 1000 individuals in the general population and is known to have a highly genetic component.  Advancements in genetic studies have been hindered by lack of consistency due to high heterogeneity of the phenotype.  It has been hypothesized that separate dimensions of the autism phenotype may influence separate genetic and biological mechanisms. Objectives: Perform a principle component analysis of all the Autism Diagnostic Interview-Revised (ADI-R) data from a sample of children with autism to draw comparison between derived factors and the existing described domains. To perform a cluster analysis of the ADI-R data to identify groups of individuals with specific phenotypic patterns that would allow sub-setting into homogenous groups for the purposes of phenotype genotype analyses. Methods: ADI-R data from 200 individuals with autism were subjected to principal component and partition cluster analysis using STATA. Results: Principal component analysis identified five main factors: verbal communication, social ever, social current, insistence on sameness, and repetitive sensory motor actions.  Cluster analysis identified two clusters from our sample representing a high functioning group and a low functioning group. Conclusions: ADI-R contains quantitative factors and there is a definable split between low and high functioning individuals.