PGC Mega-Analysis of 5300 Individuals with ASD Yields a Genome-Wide Significant Association with the Astrotactin 2 (ASTN2) Gene

Background:  Presented on behalf of the Psychiatric Genomics Consortium: Autism Spectrum Disorder Working Group. Although copy number variants and rare mutations have been demonstrated to increase risk for ASD, empirical evidence supporting a large role for inherited common variants as risk factors for ASD risk is still lacking. 

Objectives:  In order to discover common variants that increase risk for ASD, the members of the Autism Spectrum Disorders Working Group of the Psychiatric Genomics Consortium conducted a large high-density genome-wide meta-analysis (mega-analysis) combining data from multiple genome-wide association studies on 6500 individuals with ASD.

Methods:   The mega-analysis included data from six family-based studies across five different genotyping platforms. All raw genotype and phenotype data were uploaded to a central server and processed through the same quality control, imputation, and analysis procedures.  Trios were converted to cases and matched pseudo-controls using PLINK_v1.07. Imputation was performed using SHAPE-IT/IMPUTE routines benchmarked against the 1000 genomes project (build v3.Aug2012). Individuals were excluded from analyses if they were assessed at less than 36 months old or if there was any evidence contradictory to an ASD diagnosis from the ADI-R and/or ADOS.  The primary analyses were restricted to individuals of “European” ancestry, defined as having close similarity in genotype to the CEPH HapMap population (not simply geographic location). A total of 5305 individuals met diagnostic and ancestry criteria. Secondary analyses explored association with a strict autism diagnosis, cognitive ability, verbal status and gender. Association testing was carried out using logistic regression of imputed dosages. Analyses were performed using PLINK v1.07.  Fixed effects meta-analysis was performed using METAL, weighted for the inverse standard error of the effect. 

Results:   We observed a genomewide significant association for the ASTN2 SNP rs7026354 (OR=1.17; p=6.7x10^-9). Each of the contributing studies reported an association in the same direction. In addition to the significant association with ASTN2, strong associations were observed with previously implicated ASD genes, including EXT1 (rs7836146; OR=0.85; p=9.16x10^-7) and MACROD2 (rs6079556; OR=0.88; p=2.18x10^-6). In a replication study using an independent set of 1500 cases and 51K controls, we observed effects in the same-direction for 19 of the 26 SNPs passing a threshold of p < 1x10-5, a statistically significant result (p=0.014). Exploratory analyses with alternate phenotypes yielded interesting, although not statistically significant, results.

Conclusions:   In the largest genome-wide association study of ASD to date, we observed a single genome-wide significant association at the ASTN2 locus on chromosome 9.  ASTN2, a cell adhesion molecule expressed in the brain, is thought to have a role in neuronal migration.  Rare copy number deletions in ASTN2 have been observed previously in individuals with ASD. Results of this GWAS mega-analysis indicate that common variation is important and detectable in etiologic studies of ASD and contribute to genetic liability.

This work is presented on behalf of the Psychiatric Genomics Consortium: Autism Spectrum Disorder Working Group