Sponsored by:
The etiology of autism spectrum disorder (ASD) involves substantial genetic and environmental components as well as their complex interplay. One particular type of gene-environment interplay that caught many geneticists' attention is "genetic nurture": parents' genotypes could influence parental behavior and family environment which in turn affects children's phenotypes. Due to the correlation between parents' and children's genotypes, genetic effect estimates in a regular genome-wide association study (GWAS) are mixtures of direct effects (one's genotypes affecting one's phenotypes) and indirect genetic effects (one's genotypes affecting their children's phenotypes). Careful dissection of direct and indirect genetic effects has revealed strong evidence of maternal indirect effects on birth weight and cognition and have fundamentally changed our understanding of the genetic basis of these traits. Here, we employed an unconventional study design to search for evidence of indirect maternal effects on ASD risk, leveraging a large number of trios of ASD probands and healthy parents from SPARK and Simons Simplex Collection cohorts. We performed a GWAS where the mothers and fathers of ASD probands were coded as cases and controls, respectively (N = 10,496). Conditional on children being ASD cases, and assuming no indirect paternal genetic effect, this GWAS estimates the indirect maternal genetic effect on children's ASD risk. We found a suggestive association (rs79371708; P = 2.4E-7) at the TAF4 locus which is a known casual gene for ASD. Complementary to the maternal effect GWAS, we then performed a transmission disequilibrium test GWAS using the same ASD trios, which estimates the effects of transmitted variants alone and thus quantifies the contribution of direct genetic effects. Direct and maternal ASD effects had a moderate genetic correlation of 0.45 (SE = 0.2), suggesting a substantially different genetic basis. Genetic correlations between direct/maternal ASD effects and other complex traits also showed divergent patterns. We found significant correlations between maternal ASD effects and major depression, cigarette smoking, and a younger age at first birth. Instead, direct genetic effects on ASD showed significant correlations with cognition and educational attainment. These results showed compelling evidence for indirect maternal genetic effect on ASD and provided critical new insights into how self and maternal genotypes jointly influence ASD risk in children.