Finding the missing gene–environment interactions

Excerpt

Gene–environment interactions, where the biological effect of an exposure depends on an individual’s genotype, are widely held to be ubiquitous—and rightly so, considering epidemiologists have long abandoned the paradigm of ascribing disease to either “nature” or “nurture” (if indeed they ever thought of etiology in unifactoral terms) and now seek to understand the joint action of both “nature” and “nurture.” However, statistical interactions, where a quantitative measure of exposure effect differs according to genotype, are far from ubiquitous in epidemiologic studies of human disease [1]. The small number of replicated gene–environment interactions in human observational studies stands in sharp contrast to the widespread evidence for gene–environment interaction from experimental studies in model organisms [2]. This discrepancy is a puzzle. Is there something fundamentally different about the biology of human complex traits? Are there limitations to how gene–environment interactions have been studied in humans? Or both? …