# The predictive value of polygenic risk scores for depression in gene-environment interaction studies: a systematic review

## TL;DR

This review examines how polygenic risk scores for depression interact with environmental factors in predicting depression, finding mixed evidence for gene-environment interactions.

## Contribution

This is the first systematic review of polygenic risk score-based gene-environment interaction studies for depression.

## Key findings

- Most studies found significant main effects of polygenic risk scores and environmental factors on depression.
- Significant gene-environment interactions were mostly observed in large cohorts with over 40,000 participants.
- Environmental factors were often found to correlate with an individual's polygenic risk score.

## Abstract

According to the diathesis-stress model, genetic liability and environmental exposures interact in the pathogenesis of depression. Polygenic risk scores for depression (PRSD) based on large-scale genome-wide association studies have opened new avenues for investigating gene-environment interaction (GxE) beyond candidate gene studies. To the best of our knowledge, this is the first systematic review of studies that have taken a polygenic score approach to study GxE interaction effects on depression phenotypes.

Based on a preregistered, systematic literature search according to PRISMA guidelines, 56 studies were considered for qualitative analysis. Respective studies investigated a broad range of adverse and protective environmental exposures across the lifespan, e.g., trauma, stressful life events, social environments and (un)healthy lifestyle factors, using cross-sectional and longitudinal designs.

While most studies reported significant main effects of an individual’s PRSD and different environmental exposures on depression phenotypes, the overall evidence for GxE interactions was considerably heterogeneous. Findings of significant PRSDxE interactions mostly stem from large cohort studies comprising > 40000 participants, in particular, when recent environmental exposures were considered.

Two general conclusions can be drawn from this review. First, PRSDxE interactions, if at all, add a small amount of explained variance in depression phenotypes to the corresponding additive model and may thus require large samples to be reliably detected. Second, in a considerable number of studies, different environmental exposures were found to depend on an individual’s PRSD, indicating significant gene-environment correlation. We further discuss limitations, future directions and potential clinical relevance of PRSxE research in depression.