Using polygenic scores to assess liability to antisocial behavior

To date, several genome-wide association studies (GWAS) of antisocial behavior (ASB) have been conducted in Europeans, which promoted research aimed at evaluating liability to ASB-related phenotypes in independent samples. Such studies implemented a polygenic score (PGS) approach, which represents a composite score considering a number of "risky" alleles. Since no GWAS of ASB has been conducted in Russians, the present study aimed to perform a replication study of liability to severe criminal behavior (homicide) in individuals from Russia using PGS. Moreover, we sought to obtain the best model considering PGS and potential social factors as predictors. Genotyping of the "top" ten SNPs previously identified in GWAS meta-analysis of ASB (CADM2, REV3L, FOXP1, FOXP2, BDNF, FURIN, XKR6, TMEM18, SORCS3, and ZIC4 genes) was conducted via real-time PCR in 227 homicide offenders and 254 healthy donors from the Volga-Ural region of Russia. Multiple regression models included "weighted" and "unweighted" PGS and potential social factors as predictors. The best regression model of liability to severe ASB was based on genetic effects of examined SNPs and social predictors, including traumatic brain injury, severe chronic disease, and tobacco smoking, which was more pronounced among subjects with a family history of mental illness (p = 2 × 10-13). PGS alone explained a small proportion of variance in liability to ASB (1.1-1.5 %), while the inclusion of social parameters increased variance explained (16.2-21.2 %). Revealed findings evidence a higher impact of social factors than a composite effect of selected "top" SNPs in predicting liability to ASB in the examined cohort. A higher probability of ASB was linked to comorbid substance abuse, traumatic brain injury, and family history of mental illness, which may also represent a result of a "risky" genetic profile.