Background Most gene‐environment interaction studies (G × E) have focused on single candidate genes. This approach is criticized for its expectations of large effect sizes and occurrence of spurious results. We describe an approach that accounts for the polygenic nature of most psychiatric phenotypes and reduces the risk of false‐positive findings. We apply this method focusing on the role of perceived parental support, psychological control, and harsh punishment in depressive symptoms in adolescence. Methods Analyses were conducted on 982 adolescents of Caucasian origin (Mage (SD) = 13.78 (.94) years) genotyped for 4,947 SNPs in 263 genes, selected based on a literature survey. The Leuven Adolescent Perceived Parenting Scale (LAPPS) and the Parental Behavior Scale (PBS) were used to assess perceived parental psychological control, harsh punishment, and support. The Center for Epidemiologic Studies Depression Scale (CES‐D) was the outcome. We used gene‐based testing taking into account linkage disequilibrium to identify genes containing SNPs exhibiting an interaction with environmental factors yielding a p‐value per single gene. Significant results at the corrected p‐value of p < 1.90 × 10−4 were examined in an independent replication sample of Dutch adolescents (N = 1354). Results Two genes showed evidence for interaction with perceived support: GABRR1 (p = 4.62 × 10−5) and GABRR2 (p = 9.05 × 10−6). No genes interacted significantly with psychological control or harsh punishment. Gene‐based analysis was unable to confirm the interaction of GABRR1 or GABRR2 with support in the replication sample. However, for GABRR2, but not GABRR1, the correlation of the estimates between the two datasets was significant (r (46) = .32; p = .027) and a gene‐based analysis of the combined datasets supported GABRR2 × support interaction (p = 1.63 × 10−4). Conclusions We present a gene‐based method for gene–environment interactions in a polygenic context and show that genes interact differently with particular aspects of parenting. This accentuates the importance of polygenic approaches and the need to accurately assess environmental exposure in G × E.