Association between Antisocial Behavior Polygenic Scores and Adolescent Subclinical and Clinical Externalizing Behaviors

• In Event: 1-125 - Genetically-Informed Studies of Adolescent Substance Use and Psychopathology

Integrative Statement

Twin and family studies consistently demonstrate that a substantial amount of the variance in antisocial and related externalizing behaviors is attributable to genetic influences. Yet, there are few genes or genetic variants that, on their own, account for much variation in antisocial behavior. This likely reflects the polygenic nature of complex outcomes like antisocial behaviors, that are driven by many genes of typically small effects from across the genome. In an effort to capture the polygenic nature of genetic influences on antisocial behavior, we used a polygenic scoring approach to address the question of whether genetic predisposition toward antisocial behaviors are associated with genetically correlated phenotypes in adolescents. We tested whether antisocial behavior polygenic scores predicted subclinical and clinical levels of externalizing behaviors in adolescents.

The sample included 1187 adolescents (mean age = 13.98; SD = 1.82; 51% male) of European ancestry from the Collaborative Study on the Genetics of Alcoholism (COGA) Prospective Study. Subclinical levels of externalizing behaviors included Achenbach Youth Self Report (aggressive, rule breaking, externalizing), NEO conscientiousness (i.e., constraint), Barratt impulsivity, and sensation seeking. Clinical levels of externalizing included oppositional defiant disorder (ODD) and conduct disorder (CD) criteria as assessed using the Semi-Structured Assessment for the Genetics of Alcoholism (SSAGA). Antisocial behavior polygenic scores (ASB-GPS), indexing individuals’ genetic predisposition to antisocial behavior, were calculated in the COGA Prospective sample using genome-wide association estimates from the Broad Antisocial Behavior Consortium (Tielbeek et al., 2017). We conducted regression analyses to examine whether ASB-GPS were associated with subclinical and clinical measures of externalizing in adolescents. Covariates included age, sex, and two genetic ancestry principal components. We fit the regression models using SAS PROC SURVEYREG (for continuous outcomes) and PROC GLIMMIX (for binary outcomes) to account for family-based data.

Results indicated that higher ASB-GPS were associated with several subclinical externalizing dimensions: higher aggressive behavior, higher rule breaking, higher externalizing, and lower constraint (NEO conscientiousness). There was no association between ASB-GPS and impulsivity and sensation seeking. We also did not find association between ASB-GPS and clinical ODD and CD criteria. Together, our findings showed that polygenic scores derived from a recent large scale GWAS of broad spectrum antisocial behavior (pooling across multiple adult and child samples) predicted several dimensions of subclinical, but not clinical, externalizing behaviors in adolescents. Our results suggest that capitalizing on large-scale gene identification efforts and bringing those results forward to calculate polygenic scores in smaller independent developmental studies represents a useful strategy for characterizing how risk unfolds across development.

Group Authors

COGA Collaborators

Authors

• Sally I-Chun Kuo, Virginia Commonwealth University
  Presenting Author

• Jessica Salvatore, Virginia Commonwealth University
  Non-Presenting Author

• Danielle M. Dick, Virginia Commonwealth University
  Non-Presenting Author