Genome-Wide Association Meta-Analysis of Autistic Traits In The General Population

Abstract

BackgroundAutism Spectrum Disorders (ASDs) are highly heritable and recent findings indicate that at the group level, individuals with ASDs have a larger intracranial volume (ICV, a measure of total brain volume) compared to people without ASDs (results from the ENIGMA-ASD consortium comparing 1027 participants with ASDs and 1047 controls; van Rooij et al., submitted). There is now substantial evidence that ASDs represent the extreme end of a continuum of autistic traits within the general population. We previously showed shared genetic etiology and biology between autistic traits in the general population (n=1981) and ASDs (Bralten et al., Mol Psych, in press), which implies that general population studies of autistic traits could yield additional, novel ASD genes and loci. MethodsWe used a customized self-report questionnaire of autistic traits containing six items from the DSM-5 section about ASDs and twelve items from the Autism-Spectrum Quotient (AQ), administered to participants from the Nijmegen Biomedical Study (NBS), an adult population sample. We extended our previous genome-wide association study (GWAS) in NBS (new n=3098) investigating five factors/subscores - i.e., ‘childhood behaviour’, ‘rigidity’, ‘social skills’, ‘attention to detail’, and ‘imagination’ - constituting the best-fitting model to explain the observed variance in the total score of autistic traits. We are currently increasing the sample size of the autistic traits GWAS by meta-analyzing the NBS data with data from two other cohorts:(1) the Brain Imaging Genetics (Guadalupe et al. 2014) cohort (n=436), which has the same questionnaire available, and (2) the Raine study (n=965; Jones et al. 2014), which has AQ data available.We integrated the data of the ASD GWAS data from the large Psychiatric Genomics Consortium Autism Working Group (PGC-ASD; GWAS of 5305 ASD cases and 5305 pseudocontrols), with GWAS summary statistics from a recent, large-scale meta-analysis of ICV across 50 cohorts worldwide by the ENIGMA consortium (n=13171) testing potential mediating brain measures (Hibar et al. 2015). We performed LD score regression to investigate genetic correlation and SNP Effect Concordance Analysis (SECA) to investigate pleiotropy and concordance. ResultsIn total, 174 single nucleotide polymorphisms (SNPs) were found to be associated with the total autistic traits score or any of its five factors/subscores at P < 1.00E-06 in our NBS cohort. Rs141783228, a SNP within the ECT2 gene on chromosome 3, was the SNP most significantly associated with the total autistic score at P=9.3E-08.Using summary statistic data of PGC-ASD and ENIGMA-ICV we were able to show significant genetic correlation between clinical ASD and population-based ICV measures (0.27, p=0.0025) and significant pleiotropy (p=0.0015) and concordance (p=0.0033) in the expected direction (larger ICV was associated with higher ASD risk). DiscussionOur findings strengthen our hypothesis that novel ASD genes and loci can be identified by performing GWASs of autistic traits assessed through self-report questionnaires in the general population. We also showed a genetic correlation between ASDs and ICV indicating that the mechanism underlying ASD susceptibility may work through brain volume. We are excited to test how the genetic association results of our meta-analysis will stand compared to the genetics of ASDs and the genetics of an ASD-relevant brain measure (ICV), which will be presented at the conference.