Approved research

Developing and implementing new methods for evaluating the assumptions required for Mendelian randomisation

Lay summary

Mendelian randomisation is a popular method that uses genes to study the health effects of traits that are otherwise very difficult to study. It does this by making assumptions about the relationship between genes and the health characteristics being studied. It is not possible to know for certain whether these assumptions are true or not. Some checks exists, however, that can give us a better idea of whether these assumptions are true or not. When these checks are used and suggest the assumptions are likely to hold, we can sometimes have more confidence in results from studies that use Mendelian randomisation. Unfortunately, many of these checks are not used. We are proposing using one of these checks, that has never been used before in the Mendelian randomisation research, on a number of questions that have already been answered in the UK Biobank to see how confident we should be about the results. We have also developed a new way of checking how much results from Mendelian randomisation would be affected if one specific assumption was not true. This will help researchers using Mendelian randomisation understand how important this assumption is to having correct results. Together, this group of checks can provide tools for researchers to help them better gauge their confidence in their findings from Mendelian randomisation studies.