Mendelian randomization to hypothesis-free causal inference

A genetic technique called Mendelian randomization is being refined to help researchers tell whether a specific factor actually causes a disease, rather than just being associated with it. The core problem is that observational studies—where scientists simply watch what happens to people over time—are easily fooled. If people who drink coffee have less heart disease, it might be the coffee, or it might be that coffee drinkers tend to be wealthier, exercise more, or have other healthy habits. Mendelian randomization sidesteps this by using genetic variants as stand-ins for the factor of interest. Because genes are assigned randomly at conception and aren't altered by lifestyle, they act like a natural randomised trial. This programme will strengthen the reliability of that approach. The researchers will investigate how to apply it to identify treatments for existing disease—not just causes of new disease. If successful, the work could make drug development more efficient by flagging which biological targets are genuinely worth pursuing, reducing the number of clinical trials that fail because they were based on misleading observational evidence. This is primarily a methodological project: it does not aim to discover a specific new treatment, but to improve the statistical toolkit that other researchers use to find them.