GxE in Health and Well-Being

From theoretical, evolutionary, and experimental non-human genetic perspectives, gene-environment interactions (GxE) are expected to influence human behavioral outcomes. Existing approaches to modeling GxE, however, take... [ view full abstract ]

From theoretical, evolutionary, and experimental non-human genetic perspectives, gene-environment interactions (GxE) are expected to influence human behavioral outcomes. Existing approaches to modeling GxE, however, take narrowly defined approaches to selecting genes, environments, outcomes, and models, rather than seeking to develop a bird's-eye view of patterns of GxE influence. Such an omnibus perspective proved useful to developing an understanding the typical structure of genetic main effects, demonstrating that, across phenotypes: main effects are numerous and individually small; and phenotypic correlation implies genetic correlation, whereby variants identified as associated with one outcome can be used as reliable indicators of others. We can ask similar questions in a GxE space. That is, what is the likely distribution of GxE influences, in terms of both effect sizes and effect shapes; and to what extent are sources of GxE outcome-specific versus general. This project explores patterns of GxE across a variety of health-relevant outcomes (subjective well-being, neuroticism, tobacco use, and body mass index) by educational attainment (as an indicator of resource stability versus uncertainty). Patterns of heritability and prediction are compared to understand the distribution and consistency of GxE shape, structure, and size across phenotypes and modeling approaches. Estimation of effects in both twin samples and among unrelated individuals provides unique information that allows us to restrict the universe of likely effects embedded within the larger pattern of observed data. Results suggest potential for the development of higher-order theories of GxE that can be applied broadly to complex mental and physical health outcomes.