Common Variants for Schizophrenia Do Not Predict Autism

Background: The results of recent studies suggest that an increasing number of rare genetic variants can lead to both schizophrenia and autism. These findings provide new insights to the question as to whether autism ought to be considered as related to, or rather, distinct from schizophrenia.

In a recent paper published by the International Schizophrenia Consortium (ISC) evidence in support of a polygenic contribution to schizophrenia was presented. Focusing on the effect of common variants with small individual effects, they demonstrated a significant “en masse” effect of this class of genetic variation for schizophrenia (the polygene score). Interestingly, they showed that the same schizophrenia-derived polygene score also contributes to the risk of bipolar disorder.

Objectives: Given the clinical overlap between schizophrenia and ASD, as well as the molecular evidence of shared genetic risk between these disorders due to rare genomic CNVs, we hypothesized that common risk alleles may also be shared between schizophrenia and autism.

Methods: To test this hypothesis, we utilized the polygenic score which was derived from a schizophrenia case-control dataset, (previously reported by the ISC).  We examined whether this schizophrenia-derived polygenic score was able to differentiate autism cases from controls (based on trio data available through the Autism Genome Project, AGP).

Results: For the current analysis 2,737 proband-parent trios from the AGP were included. In previous publications methods of phenotyping of these samples have been extensively described. Of all samples high resolution SNP data (Illumina 1M) were available.  Given the absence of unrelated controls in the AGP dataset, the non-transmitted alleles to the proband were used as controls (or pseudocontrols).  We used the schizophrenia score allele set that was previously described in the ISC paper, consisting of ~74k SNPs with nominal association (at different significance thresholds) in a discovery GWAS schizophrenia case control sample.

Our results show that the schizophrenia-derived polygenic score was not significantly different between ASD-cases and controls.

Conclusions:  

While there may be a number of rare causative genetic variants that are shared between schizophrenia and autism, this study suggests that such sharing is non existent (or very limited) when common genetic variants are considered. These findings, in the context of other recent findings provide important novel insights into shared and distinct elements of the genetic architecture of autism and schizophrenia.