Alternative models for the contribution of psychosocial stressors and genetic susceptibility to the risk of depressive symptoms. From Fuller et al. (2021).
Collaborators and I have a new paper just out in the American Journal of Physical Anthropology, based on data from the HEAT Heart Health Study. The central question is to what extent variation in depressive symptoms is associated with individual-level genetic differences and exposure to psychosocial stressors. We tested two alternative models, shown at right.
Model (a) says that psychosocial stressors are primary contributors to the risk of depressive symptoms, but that individual-level genetic differences may make some people more sensitive to psychosocial stressors. Model (b) says that genetic susceptibility and psychosocial stressors contribute independently to the risk of depressive symptoms. The thickness of the lines in these diagrams is meant to illustrate expectations about the relative magnitude of each association.
Contrary to some prior theory, our results fit model (b), not (a). We tested five single-nucleotide polymorphisms (SNPs) that had been associated with depressive symptoms in other studies. None was associated with CES-D in our sample. By contrast, a regression model including two psychosocial stressors (difficulty paying bills and the proportion of people in one’s social network who are a source of stress or worry) accounted for 17% of the variance in depressive symptoms. When we tested psychosocial stressors and genetic susceptibility variants simultaneously, one SNP (rs1360780 inFKBP5) was associated with CES-D—but only when psychosocial stressors were in the model.
This paper adds to the weight of evidence pointing to the primacy of social context in the risk of depressive symptoms. It also suggests that genetic influences on depression-related phenotypes can’t be understood without attention to the social and material conditions in which people live. By extension, improving measurement of those conditions may empower genetic association studies.
