Autism spectrum disorders (ASD) are neurodevelopmental in origin and characterized by impairments in reciprocal social interaction and communication accompanied by restricted and repetitive patterns of interest or behavior. With improved surveillance and a broadening of the diagnostic criteria, the most recent prevalence studies suggest that ASD may affect as many as 1/166 children in the US. Treatments are limited resulting in little impact on the profound morbidity. Little is known about the etiology of ASD, but there is a strong genetic component. Genetic studies over the past decade have failed to identify definitive ASD genes, but have clearly shown that the underlying genetics are more complex than anticipated with the likelihood that multiple genes are acting independently and/or interactively. With this realization the field of ASD genetics is at a critical juncture. To move forward we must embrace new and creative paradigms to successfully dissect the genetic etiology of this disease. Specifically the current study will:

1. Extend our ascertainment to the full range of ASD. We will increase our ASD dataset by collecting 600 families who meet clinical diagnostic criteria for the three most common ASDs: autism (AUT), Asperger disorder (ASP) and Pervasive Developmental Disorder-NOS (PDD-NOS). Phenotype assessment will be expanded to include the Social Responsiveness Scale (SRS) and detailed seizure data. Ascertainment will focus on parent-child trios.
2. Further characterize the ASD-associated variations in the Gamma-Aminobutyric Acid (GABA) receptor (GABR) subunits. We will complete our GABR subunit variant analysis, including coding, UTR, regulatory, and conserved intronic sequence, concentrating on GABRA4 and GABRB1. We will also investigate the role of variation in other GABR subunits. All AUT specific findings will be examined in the entire ASD dataset.
3. Identify the chromosome 19p ASD gene. Multiple genome-wide screens have provided significant evidence for linkage to Chromosome 19p and we have found evidence for significant association in the linked families within an ~3 Mb region. We will genotype 768 SNPs spaced at ~4 kb to refine the candidate region and delineate specific candidate genes for further testing. We will screen the strongest associated candidate genes for additional variants and evaluate evidence for association in the entire ASD dataset (SA5).
4. Develop and apply the Phenotypic Homogeneity Distinction (PHD) algorithm for identification of the phenotypic signature of linkage and/or association signals. The PHD algorithm identifies phenotypic covariates that discriminate a target sample subset (e.g. the positively linked families or families carrying the associated risk allele) from its complement. The resulting PHD discriminators will define a homogeneous subset from our newly-ascertained ASD dataset to use, as a confirmatory dataset and for localization within genes under study in our other aims.
5. Test for gene-gene interactions. To fully explain the spectrum of autism risk, we will test for interactions between genetic variations associated with ASD (such as those in GABR subunit genes). We will build upon our experience using the MDR method and extend our analyses across biological pathways using the entire ASD data set.