Evaluating the Polygenetic Susceptibility to Peer Reported Bullying Behaviors

Bullying and peer victimization have been a topic of national attention given recent violence in the schools.  Although environmental risk factors for bullying have been well characterized (Swearer et al., 2010), genetic influences on bullying are not well understood.  Recent work, however, has begun to examine potential genetic influences on bullying.  Twin studies reveal a complex interplay between environment and genetics is likely (Ball et al., 2008).  However, genetic susceptibility to bullying, like other complex traits, may be polygenic in nature, such that consideration of a genetic risk score that summarizes the underlying polygenetic susceptibility of the trait can be useful.   Drawing upon longitudinal data, this presentation explores the role of genetics on early childhood bullying behavior as measured by peer-report from the fall of first grade.  Use of a polygenic score derived from a conduct disorder genome wide association study (GWAS) will allow us to explore the role of genetics on these behaviors across time, as well as explore any potential moderators such as a classroom-based intervention.  Data used for this study come from a community based trial in an urban school district beginning in 1993 (Ialongo et al., 1999).  Data come from 678 urban, primarily African American children followed from first grade through young adulthood.  Peer nominations of bullying behavior were taken from the Peer Assessment Inventory (PAI; Dolan et al., 1993).  During the 19-21 year follow-up assessment blood or saliva was collected from participants and DNA was genotyped using the Affymetrix 6.0 microarray.  This microarray provides approximately 900,000 SNPs from across the genome.  SNPs from the rest of the genome were then imputed.  Using the SNP list from a GWAS focused on conduct disorder, a polygenic risk score (PRS) was created from the SNPs reaching a threshold of significance.  Peer nominations of early bullying behaviors were modeled using latent profile analysis and demonstrated three profiles: low involvement in bullying (72%), moderate levels of bullying behaviors (23%), bully-victim profile (5%), which was characterized by moderate levels of bullying behaviors and victimization.  Results suggest that the PRS was significantly associated with class membership, with individuals in the bully-victim profile having the highest levels of the PRS.  Additional analyses will be conducted to explore variation in these associations by gender. This line of research has important implications for understanding variation in preventive interventions in relation to genetic factors.