Mendelian Randomization (MR) is suitable to investigate gene - environment interactions when certain genetic variants (i.e. SNPs) are found to be strongly associated with the exposure. Examples in the context of metabolomics:
- SNPs => enzymes => metabolites
- SNPs => gut microbiome compoistion => metabolites
- SNPs => dietary preference => metabolites
MR leverages the fact that individuals naturally inherit different “doses” of specific genetic variants (e.g., 0, 1, or 2 copies of a particular allele). These “doses” of genetic variants often lead to different average levels of exposure within the population. Because these genetic variants are randomly allocated at conception, their “dose” is largely independent of environmental and lifestyle factors that could otherwise confound observational studies. MR then assesses whether these genetically determined differences in exposure levels (the “treatment” effect induced by the genetic “dose”) are associated with differences in the outcome. If they are, and the core MR assumptions hold, it provides strong evidence for a causal link.