Data mining the impact of blood metabolites on risk of disease

Rationale
Blood metabolites are associated with a range of phenotypes and disease outcomes [cites]. A number of studies have also explored the causal role of specific metabolites on disease outcomes such as coronary heart disease using Mendelian randomization [cites]. However, both new data and novel methods now enable us to explore the causal role of metabolites across an extensive range of phenotypes and disease outcomes, gaining new insights into their funcitonal role in health.
The UK Biobank [1] offers a wealth of genetic and phenotypic data, enabling a wide range of epidemiological hypotheses to be explored. In parallel to this we have implemented a novel pipeline that applies Mendelian randomization [2] across a wide range of UK Biobank disease outcomes to estimate the causal effects of any heritable trait.

Aims & objectives
The aim of the project is to explore the causal role of the blood metabolome on risk of a range of diseases. You will apply existing methods and develop new approaches to dissect the complex relationships between metabolites and a range of disease outcomes.

Methods
This data mining project will be primarily computational, and will enable you to develop skills in programming, statistics and epidemiology. Methods include: - Generate weighted allele scores for metabolites using data from all published metabolite GWAS and implement these in UK Biobank - Perform two-sample Mendelian randomization to identify the network of causal relationships between metabolites and a wide range of phenotypes and diseases using UK Biobank - Develop statistical and bioinformatic approaches to resolving correlated effects amongst correlated metabolites - Construct a network of causal relationships between metabolites and disease outcomes.

When applying please select ’PhD in Social and Community Medicine’ from the Faculty of Health Sciences.