Fig. 1

a Illustration of Mendelian randomization to study causality between exposure and outcome in the presence of confounders. b Illustration of a situation when the assumptions of Mendelian randomization are violated because the genetic instrument has pleiotropic effects, influencing both the exposure and the outcome. c The assumptions are violated because the genetic instrument is in linkage disequilibrium with another variant that is associated with the outcome. d The genetic instrument is associated with confounding factors, violating the assumptions of Mendelian randomization. X, modifiable exposure of interest; Y, outcome of interest; Z, genetic variant (a single allele or a linear combination of several alleles) used as an instrumental variable; U, (measured or unmeasured) confounding factors; G, genetic variant that is in linkage disequilibrium with Z. The broken arrow denotes an assumed, but unknown, causal effect