Fig. 1From: No evidence of genetic causality between diabetes and osteonecrosis: a bidirectional two-sample Mendelian randomization analysisThree key assumptions of the MR study : A (â‘ ) Relevance assumptions: SNPs are strongly associated with exposure; B (â‘¡)Independence assumptions: SNPs are independent of confounders; C (â‘¢) Exclusivity assumption: SNPs must only affect outcome via exposure. SNPs single-nucleotide polymorphismBack to article page