GxE interactions in child psychopathology, findings from DREAM BIG consortium

Context: Previous DREAM BIG consortium work demonstrated the main effect of various polygenic scores (PGS) and prenatal maternal affective symptoms on child psychopathology. However, investigating how genes and the environment (GxE) interact is crucial to understanding psychopathology across development. Developing integrative models with effects significant enough to lead to clinical and policy changes requires systematic replication. Accordingly, in DREAM BIG, we examine whether PGSs interact with perinatal adversity (maternal affective symptoms and maternal socio-environmental adversity) to predict three aspects of child psychopathology (general psychopathology, and internalizing and externalizing symptoms). Methods: Using the harmonized ALSPAC (UK), GEN-R (Netherlands), and MAVAN (Canada) cohorts, we used Gеnеralizеd Random Forеst to analyze GxE intеractions and meta-analyzed the obtained estimates. This methodology is ideal for investigating interactions between PGSs and environmental variables as it enables a nuanced estimation of heterogeneous effects. Results: Significant GxE interactions emerged. For general child psychopathology, four PGSs (schizophrenia (p = 0.01), insomnia (p = 0.04), autism (p = 0.02), and phobia (p = 0.01)) x socio-environmental adversity and one PGS (neuroticism p = 0.017) x maternal affective symptoms (M-factor) interactions were observed. For internalizing child psychopathology, three PGS (ADHD (p = 0.021), autism (p = 0.015), and cross disorder (p = 0.031)) x socio-environmental adversity and three PGSs PGS (autism (p = 0.019), panic (p = 0.002), and phobia (p = 0.005)) x M-factor interactions were observed. For externalizing child psychopathology, two PGS (cross disorder (p = 0.086) and alcohol (p = 0.095)) x socio-environmental adversity and four PGS (alcohol (p = 0.001), cross disorder (p = 0.02), bipolarity (p = 0.027), and phobia (p = 0.102)) x M-factor interactions were observed. Conclusion: These findings extend our previous work by suggesting that while some PGSs (e.g., autism, phobia, and cross disorder) show significance across different environmental contexts (maternal affective symptoms and maternal socio-environmental adversity), others are more environment-specific (e.g., schizophrenia or socio-environmental and neuroticism for maternal mood). This highlights the complexity of gene-by-environment interactions in child psychopathology, where both the type of perinatal adversity and the specific genetic predispositions play distinct roles in shaping different psychopathological outcomes. Validation is ongoing using the Norwegian MoBa cohort.