Fig. 1

A flowchart illustrating the study design process of Mendelian randomization. (a) The selected SNPs must be strongly related to the exposure; (b) The included genetic variant must be associated with the risk of the outcome solely through the exposure, not by confounding factors; (c) The SNPs should be independent of confounders. IVW: Inverse-variance weighted; WMn: Weighted median; WM: Weighted mode; LOO analysis: Leave-one-out sensitivity analysis