Polygenic risk for aggressive behaviour from late childhood through early adulthood

• In Event: Poster Session 07
  In Poster Session: PS 07 Topic - Biological Processes: Neuroscience and Genetics

Abstract

Aggressive behavior is a common symptom of childhood psychopathology and linked to lower academic performance and social functioning (Campbell et al., 2006; Vuoksimaa et al., 2020), involvement with crime (Schaeffer et al., 2003), higher likelihood of substance use (Timmermans et al., 2008), and lower earnings (Vergunst et al., 2019). Negative outcomes are particularly common in individuals who display persistent high levels of aggression across childhood and adolescence (Cleverley et al., 2012; Reef et al., 2011). The aetiology of aggressive behavior has been studied extensively in longitudinal and genetically-informed designs, alluding to a substantial role for genes to explain individual differences in aggression across development (Hudziak et al., 2003; Luningham et al., 2020; Rhee & Waldman, 2002). This field has thus far mostly relied on twin samples to estimate the amount of variance in aggression that is explained by genetic versus environmental factors. Extending this research, we utilized a polygenic risk score (PRS) derived from a genome-wide association study into child aggression (PRSAGG, Pappa et al., 2015) to analyze genetic influence on the development of aggression from late childhood through early adulthood. Data come from the first six waves of the Dutch TRacking Adolescents' Individual Lives Survey (TRAILS), where the first wave took place when children were ~11 years old and the sixth wave took place at ~25 years. Aggression was assessed from parents (waves 1-3, wave 5), participants (all waves), and teachers (waves 1-3).
Figures 1a-1c depict results from latent class growth models. For parent reported aggression, the 3-group model fit the data best whereas for self- and teacher-reports 4-class models fit better. For all reporters, the largest group showed low levels of aggression symptoms across development and around 1 in 10 individuals showed persistent high levels of aggression.
Genetic data were available for a subset of participants (n = 1354). For those, correlations with aggression items and multinomial logistic regressions with trajectory class as outcome were computed. Statistically significant associations were found for all parent-reported aggression items as well as self-reported aggression at wave 3 and teacher-reported aggression at waves 1 and 3. Trajectories based on parent-report were predicted by PRSAGG such that higher genetic risk for aggression decreased the likelihood for being in the Low compared to the Decreasing and High class. No associations were found between PRSAGG and self- or teacher-reported aggression trajectories.
In sum, stable associations with the PRSAGG appeared for parent-reported aggression only. For those, the PRS-based approach taken in this study supports findings from quantitative genetic research, namely that genetic factors are partly responsible for individual variation in the development of aggression. However, it is also possible that the predictive strength of the PRSAGG is higher for constructs that are more similar to those on which the GWAS from which the summary statistics were draw are based. Replications of the analyses presented here are needed to shed light on these questions.

Authors

• Isabelle Ouellet-Morin, Université de Montréal
  Non-Presenting Author

• Tina Kretschmer, University of Groningen
  Presenting Author

• Charlotte Vrijen, Rijksuniversiteit Groningen (UG/RUG)
  Non-Presenting Author

• Catharina Hartmann, Accare
  Non-Presenting Author