Skip to main content
. 2021 Mar 15;9:643582. doi: 10.3389/fcell.2021.643582

FIGURE 1.

FIGURE 1

A directed acyclic graph of MR model. IV assumption 1: IVs are strongly correlated with exposure. IV assumption 2: IVs are independent of outcomes (i.e., IVs can only affect outcomes through exposure). IV assumption 3: IVs are not related to confounding factors. MR, Mendelian randomization; IVs, instrumental variables.