Developing Mendelian randomization methodology to understand the role of obesity measures on physical and mental health - PhD in Medical Studies (Research England DTP)

Project Description

This studentship will develop novel methods for Mendelian randomization to address important scientific questions on the relationship between obesity and health. The student will develop strong quantitative skills in Statistics, extensive knowledge of genetics, epidemiology and population health.

Mendelian randomization (MR) provides a general framework for testing whether a modifiable exposure exerts a causal effect on the severity or risk of a disease, using genetic variants as instrumental variables. MR studies have helped to prove a definitive causal link between Body Mass Index (BMI) and poor health, but BMI is itself known to be a crude measure of adiposity.

Recent research is revealing a more complicated picture, with evidence that subcutaneous and visceral fat exert independent and often opposite effects on physical and mental health. Sex hormone secretion is also known to play an important role in modulating the distribution of fat in individuals over the life course. However, the extent to which these factors can be separated within an MR analysis remains unclear.

This PhD will develop novel methods for Mendelian randomization to address important questions in this scientific area. Specific areas of focus will include:

- Developing algorithms to detect and exploit distinct genetic clusters within an MR model;
- Proposing Multivariable MR models to estimate the joint effects of several adiposity measures on disease risk;
- Developing techniques to adjust for medication use within MR analyses;
- The use of Bayesian methods to account for model uncertainty and horizontal pleiotropy within an MR analysis.

Methods will be applied within large population cohort studies such as the UK and China Kadoorie Biobanks. Comparing results across different ancestries, and the extent to which causal relationships translate, will be an important aspect of the project.

The student should have a solid grounding in Statistics and experience in handling and analysing medical data. The student will join a vibrant, diverse and world class interdisciplinary research team, including geneticists, statisticians, biological scientists and epidemiologists