Fig. 1From: Associations of the circulating levels of cytokines with the risk of myeloproliferative neoplasms: a bidirectional mendelian-randomization studyStudy design overview and assumptions for MR Design. Assumption 1: IVs were not related to Confounders; Assumntion2: a strong correlation between iVs and exposure; Assumption 3: IVs can affect outcomes only through exposure and not through other pathwaysBack to article page