MR Dictionary

MR-Egger regression and extensions

A method developed for two-sample MR settings that combines Wald ratio (or ratio estimates) together into a meta-regression (with an intercept and slope parameter) to estimate the causal effect adjusted for any directional pleiotropy

All single nucleotide polymorphisms (SNPs) being used as instrumental variables (IVs) can be invalid due to pleiotropy (specifically, horizontal pleiotropy) as long as the pleiotropy satisfies the Instrument Strength Independent of Direct Effect – InSIDE - assumption. This method can be implemented in a Bayesian framework but standard implementation requires the associations between the IVs and exposure to be orientated in a positive direction. The radial MR formulation avoids this step. The intercept of a MR-Egger regression provides an indication of horizontal pleiotropy when it is not null. Similar to the inverse variance weighted (IVW) method, the NO Measurement Error (NOME) assumption (i.e., that the association between the SNP and exposure is known, rather than estimated, with no measurement error), still holds and, if violated, there may be weak instrument bias. This can be tested with the I-squared statistic with the MR-Egger method. 

References

Other terms in 'Pleiotropy-robust two-sample MR methods':