Mendelian randomisation at 20 years: how can it avoid hubris, while achieving more?

Mendelian randomisation, the initial extended formulation of which appeared 20 years ago,1 uses germline genetic variation to strengthen causal inference regarding modifiable risk factors for disease. In the original formulation, a genetic variant of known function, often a single-nucleotide polymorphism (SNP), was taken to proxy for the exposure. Mendelian randomisation was introduced at a time when it was becoming clear that conventional observational epidemiological studies could produce misleading findings, with randomised controlled trials of the exposures they identified as likely causes of disease yielding null results.