This research project aims to use the wealth of biomedical data from the UK Biobank project to better understand the causes and consequences of disease. A paramount challenge faced by researchers is how to analyse these types of data whilst avoiding sources of bias that may result in misleading conclusions. A major source of bias in this regard is known as confounding, which may lead to a lifestyle risk factor appearing to influence disease risk, even though this is in fact due to the unanticipated influence of another (confounding) factor which has not been taken into account.
This project will therefore leverage human genetic data to improve causal inference using a technique known as Mendelian randomization (MR). MR capitalises on the fact that genetic variation is randomly inherited at conception from an individual’s parents and typically remains unchanged throughout life. As such, particular genetic variants can be used to investigate how modifiable risk factors influence disease which reduces the impact of sources of bias such as confounding. For example, genetic variants which are robustly known to increase appetite could be used to investigate how having a larger body size increases risk of heart disease.
In this project we aim to develop and apply methods that will allow more reliable insight into how different modifiable lifestyle and physiological factors influence disease risk, as well as highlighting factors which to date have likely been implicated as causal through unreliable analyses. This will include applying current methods, but also involve developing novel approaches in this field. For example, there are currently a paucity of approaches that can provide reliable insight into why diseases progress faster in some people than others once they have developed. Similarly, the optimal way to utilise genetic variants as mimics of drug targets to better understand which medications may provide optimal treatment for disease requires developing.