Autism spectrum disorders are neurodevelopmental disorders impacted by genetic factors. Traditional approaches, such as genome-wide common and rare variants association analyses, provide ways to identify autism candidate regions. However, identification of susceptibility genes within these regions is difficult. Based on the knowledge that genetic variants may affect expression levels of associated or nearby genes, we map single nucleotide polymorphisms (SNPs) that are associated with expression of an autism candidate gene. These SNPs or expression quantitative trait loci (eQTLs) represent putative autism risk pathways. By identifying a regulatory network for a given candidate gene or region (based on previous SNP or CNV GWAS), we will obtain a better understanding of the underlying neurobiology and genetic architecture of autism and identify potential therapeutic targets at key regulators.

More recently, we are using eQTL data to explore how the CNVs at 16p11.2 and 22q11.2 are contributing to behavioral disorders (autism, schizophrenia, obesity, etc.). Copy number variation often leads to changes in gene expression, much like eQTLs. By using novel computational methodology (i.e. PrediXcan, MetaXcan), we can use eQTL information  to determine how changes in expression of CNV genes (in people who have normal copy number) associate with psychiatric disease.