MR-GxE is an extension to two-stage least squares (TSLS) within a one-sample MR setting that exploits the interactions between genes and covariates to test for correct for invalid instrumental variables (IVs) within a linear interaction model framework.
Requires a strong gene-environment interaction and the pleiotropic effect to be constant across covariate subgroups. The MR-GxE framework is also useful for testing the validity of MR assumptions. For example, consider estimating the causal effect between heaviness of smoking with cardiovascular disease with MR analyses. The well characterised genetic variant in the CHRNA5 gene that influences the heaviness of smoking may be a good IV candidate for this question but it is worth knowing that the variant only relates to heaviness of smoking in individuals who smoke. Any association between that genetic variant and cardiovascular disease in individuals who do not smoke indicates that the genetic variant is associated with the outcome (cardiovascular disease) independent of the exposure (smoking) and, hence, likely violates one of the key MR assumptions.
References
- Freathy RM, Kazeem GR, Morris RW, et al. Genetic variation at CHRNA5-CHRNA3-CHRNB4 interacts with smoking status to influence body mass index. Int J Epidemiol 2011; 40: 1617-28.
- Slichter D. Testing Instrument Validity and Identification with Invalid Instruments. http://wwwsole-joleorg/14436pdf 2014.
- Cho Y, Shin S-Y, Won S, Relton CL, Davey Smith G, Shin M-J. Alcohol intake and cardiovascular risk factors: A Mendelian randomisation study. Scientific Reports 2015; 5: 18422.
- Spiller W, Slichter D, Bowden J, Davey Smith G. Detecting and correcting for bias in Mendelian randomization analyses using Gene-by-Environment interactions. International Journal of Epidemiology 2018; 48: 702-712.