Gene–Physical Activity Interplay in Depression: Candidate–Gene Interactions, Polygenic Susceptibility, Lifestyle Context, and Mendelian Randomization Evidence—Systematic Review

Background/Objectives: Depression is a heterogeneous disorder shaped by both inherited liability and environmental exposures. Physical activity is a scalable, modifiable behavior consistently linked to lower depressive symptoms and reduced incident depression, but interpretation is complicated by measurement error, genetic confounding, and bidirectional pathways in which depression can also reduce activity. This systematic review synthesizes evidence on gene–physical activity interplay in depression across four complementary frameworks: (i) candidate–gene interaction studies, (ii) genome-wide susceptibility indexed by depression polygenic risk scores (PRS), (iii) lifestyle-context and activity-architecture analyses (e.g., timing and accumulation patterns), and (iv) Mendelian randomization (MR) studies testing bidirectional causal effects between activity-related traits and depression. Methods: A PRISMA-aligned search and narrative synthesis were conducted due to substantial heterogeneity in populations, exposure measurement (self-report vs. accelerometer), genetic approaches, and depression phenotypes. Twenty-seven studies met inclusion criteria. Results: Across designs, the most consistent pattern was that higher physical activity (or lower inactivity) tracked with lower depression risk or symptom burden even when genome-wide genetic susceptibility was modeled, supporting largely additive contributions of PRS and activity rather than strong, generalizable PRS × activity interactions. MR evidence most consistently supported a protective effect of physical activity on depression when activity was indexed by accelerometer-derived phenotypes, whereas self-reported activity instruments yielded weaker or more heterogeneous findings. Bidirectional genetic evidence also indicated that depression liability can causally suppress physical activity, consistent with a feedback loop relevant for prevention and intervention. Candidate-gene moderation effects were mixed and typically emerged only in specific contexts (e.g., stress history, developmental stage, sex, or treatment setting), underscoring limited replicability and sensitivity to how activity is operationalized. Conclusions: Overall, the literature supports physical activity as broadly protective across levels of genetic risk, while emphasizing that robust inference depends on objective exposure measurement, careful handling of confounding and reverse causation, and improved generalizability beyond predominantly European-ancestry genetic resources.