Leveraging large-scale molecular datasets to dissect immune-mediated pathways to disease and guide therapeutic opportunities

Abstract

Approaches that improve therapeutic target validation are essential given the high attrition rate in drug development. Human genetics offers a powerful framework to de-risk drug discovery and repurposing, with genetically supported targets more likely to succeed in clinical trials. This thesis develops and applies a genetically informed framework for drug development across immune-mediated phenotypes, integrating data from genome wide association studies (GWAS), Mendelian randomisation (MR), colocalisation, heritability analyses and protein interaction networks.

In the first case study, proteome-wide MR was applied to evaluate the causal effects of blood plasma proteins on multiple sclerosis (MS). Sixteen proteins were prioritised, some of which (STAT3, CD59, AHSG, FCRL3) colocalised with MS loci, providing high-confidence evidence of involvement in disease biology. Secondary analyses across MS comorbidities revealed limited pleiotropic effects, suggesting that most prioritised proteins act predominantly within MS. This supports their translational value as therapeutic targets while highlighting the need for refined approaches to capture shared biology across related traits.

The second case study examined the immunogenetic relationship between asthma and lung cancer. Local genetic correlation and colocalisation were applied to focus MR on shared heritable features, identifying overlapping MHC regions. However, MR analyses demonstrated divergent protein effects, in particular for TNF and IL6. These findings emphasise how underlying molecular mechanisms may diverge across related traits, with distinct therapeutic implications.

Finally, a heritability-informed repurposing framework was implemented across a broader set of immune traits by integrating MR with local heritability analysis, genetic correlation, and pharmacological annotation. Overall, 133 repurposing suggestions are identified across targets with convergent effects across traits underscored by shared genetic architecture.

Together, this work demonstrates the utility of human genetics for therapeutic prioritisation and repurposing. It contributes a generalisable framework that identifies promising targets while revealing context-specific biology, providing a systematic approach to accelerate and de-risk drug development.

Date of Award	17 Mar 2026
Original language	English
Awarding Institution	University of Bristol
Supervisor	Tom R Gaunt (Supervisor), Venexia M Walker (Supervisor) & Tom G Richardson (Supervisor)