Figure 1.
The principles of Mendelian randomization and key MR assumptions which are: (i) the genetic instrument (Z) is robustly related to the risk factor of interest (X); (ii) Z is not associated with confounders (C) of the X-outcome (Y) association; and (iii) there is no path from Z to Y other than through X (part a). Part b shows how this may be violated.
