Skip to main content
. 2014 Mar 4;22(4):186–189. doi: 10.1007/s12471-014-0534-z

Fig. 1.

Fig. 1

Causal relationships which satisfy the core assumptions of Mendelian randomisation: (1) genotype is associated with phenotype, (2) genotype is independent of confounding factors, and (3) genotype is associated with outcome, but only through phenotype