Autism spectrum disorder (ASD) is a neurodevelopmental disorder with a complex genetic etiology. Little is known about the role of inherited common variants in the development of ASD and it’s persistent, and perhaps increased, frequency in our population, despite decreased fecundity in ASD individuals. One possible explanation for this phenomenon is assortative mating. Individuals with elevated, but sub-diagnostic ASD behaviors may be more likely than chance to mate, and thus increase the likelihood of their offspring of inheriting an underlying genetic load that pushes them over the ASD diagnostic threshold. In parents of children with ASD, we investigate 1) If genetic risk for ASD, schizophrenia (SCZ) or height can predict ASD-like behaviors 2) If they are more similar than chance in their ASD-like behaviors and their genetic risk for ASD, SCZ or height and 3) If there is a correlation between ASD behavioral and genetic similarity. Analyses were performed on 2727 Caucasian parental pairs of children with ASD in the Simons Simplex Collection. All parents completed the Broad Autism Phenotype Questionnaire (BAPQ) and were genotyped via microarray. After standard quality control measures, Polygenic Risk Scores (PRS) were calculated in PLINK using summary data from published GWAS for SCZ, ASD, and height. Linear regression was performed in R to predict individual PRS based on individual BAPQ scores for all phenotypes. Degree of correlation between mothers and fathers was calculated for both BAPQ scores and PRS for SCZ, ASD, and height. Finally, the degree of correlation was calculated between the squared difference of mother and father BAPQ scores and the squared difference between mother and father PRS for all phenotypes. For parents of children with ASD, individual PRS for ASD, SCZ or height did not predict individual BAPQ total or subscale scores (b = -0.06 – 0.3 , p > 0.01). Total BAPQ score and BAPQ pragmatic subscale scores were strongly correlated between parents (r = 0.14 , p = 1.4e-12 ; r = 0.11 , p = 2.4e-8). PRS for both ASD and SCZ were also strongly correlated between parents (r = 0.16, p = 1.1e-9; r = 0.19, p = 3.9e-14). Height PRS were more weakly correlated (r = 0.08, p = 0.003). Squared difference of parental PRS scores for all phenotypes were not predictive of the squared difference of parental total BAPQ scores or any subscales (b = -0.07 – 0.02, p ≥ 0.01). Assortative mating is an established phenomenon in individuals with ASD. Here, we show that parents of individuals with ASD are also more likely than chance to highly correlate on broad autism phenotype measures, primarily on pragmatic language measures. Pragmatic language measures are, for example, when someone says inappropriate or unrelated things during conversations or has little variety in language use. Further, we found that parents of children with ASD have highly correlated PRS for both ASD and SCZ, but not height (our non-psychiatric control phenotype). Therefore, we hypothesized that the degree of correlation between ASD and SCZ genetic risk may predict correlation on total BAPQ measures or on the BAPQ pragmatic language subscale. However, the results of this analysis found that there was no significant correlation between these measures. This suggests that the genetic similarity we captured through PRS may reflect assortative mating in parents of ASD individuals due to factors other than autistic-like behaviors and that there are yet unknown factors underlying BAPQ similarities between parents of ASD individuals.
Read full abstract