Autism Polygenic Scores Are Associated with Trauma and Self-Harm

Poster Presentation
Friday, May 3, 2019: 11:30 AM-1:30 PM
Room: 710 (Palais des congres de Montreal)
V. Warrier1 and S. Baron-Cohen2, (1)University of Cambridge, Cambridge, United Kingdom, (2)Autism Research Centre, Department of Psychiatry, University of Cambridge, Cambridge, United Kingdom

Autistic individuals experience significantly elevated rates of childhood trauma, and self-harm and suicidal behaviour and ideation (SSBI). It is unclear if polygenic scores for autism are associated the likelihood of experiencing these adverse mental health outcomes, and variables that may mediate and moderate this.


To investigate if polygenic scores for autism are associated childhood trauma and SSBI and identify variables that mediate and moderate these effects.

Methods: Using data from the UK Biobank (105,222 < N < 105,638), we applied polygenic scores for autism, and tested for association with childhood traumatic events and SSBI. The exposure variables were polygenic scores for autism, derived from an independent GWAS of autism. Sex and childhood trauma were moderators. Depression and anxiety symptoms, social relationships and job satisfaction were mediators. The main outcomes were childhood trauma scores, self-harm ideation scores, and self-harm scores.

Results: Polygenic scores for autism were significantly associated with childhood trauma scores (max R2 = 0.083%, P < 2x10-16), and self-harm ideation scores (max R2 = 0.098%, P < 2x10-16), and self-harm scores (max R2 = 0.12%, P < 2x10-16). Male sex significantly negatively moderated the effect of polygenic scores on childhood trauma scores (Beta = -0.025±0.005, P = 1.59x10-5) and self-harm scores (Beta = -0.015±0.005, P = 0.007). Childhood trauma scores significantly positively moderated the effect of polygenic scores on self-harm ideation scores (Beta = 0.007±0.002; P = 0.008). Depressive symptoms, and quality and frequency of social interactions were significant mediators of the effect of polygenic scores on autism, with the proportion of effect mediated ranging from 0.23 (95% CI: 0.34 – 0.12) for depression to 0.02 (95% CI: 0.03 – 0.01) for frequency of social interaction.

Conclusions: Our results suggest the risk for SSBI and childhood trauma is partly associated with polygenic scores for autism and point to significant gene-by-environment interactions, with sex and childhood traumatic experiences moderating the effects of SSBI. Finally, our results also identify significant mediators of the effect of polygenic scores on SSBI.

See more of: Statistical Genetics
See more of: Statistical Genetics