Genetic colocalisation across disease to identify drug repurposing candidates and risk of patient co-morbidities

Background

Discoverying new drugs is expensive, time intensive and fraught with failure at all stages (1). An attractive alternative, with much lower costs and faster development timelines, is to find new applications for already approved drugs, a process known as drug repurposing or drug reallocation. Genetic evidence increases the likelihood of successful drug application (2). Therefore, another strategy to reduce drug development costs and timelines is to link genetic association results with expression data to find potential drug targets.

Pleiotropy – whereby a single gene can affect multiple traits – is widespread throughout the human genome. By studying pleiotropy of genetic signals both strategies can be combined. First elucidate which disease mechanisms are shared between diseases and then, combining the pleiotropic signals with expression data, identify novel drug targets and candidates for drug repurposing. Additionally, identifying pleiotropy between diseases allows for identification of genetic variants that have already been discovered in association with other disease phenotypes but may impact the outcome of another one, increasing the power of genetic analysis. An example of the integration of genetic and expression data to repurpose a drug indentifying a shared mechanism with another disease was the discovery of Bariticinib as a new and effective drug target for Covid-19 during the Covid-19 pandemic (3).

The aim of this project is to combine genetic studies of immunological, infectious, cardiovascular and neurological diseases in order to find shared disease mechanisms between seemingly distant diseases. Then, combine this results with expression quantitative loci (eQTLs, pQTLs) to identify drug targets for the mechanisms and therefore new candidates for drug repurposing and targeting. And finally, use the shared genetic loci to construct machine learning models for stratifying patients with risk of co-morbidities.

To perform this project we will use publicly available summary statistics from large Biobanks such as UK Biobank and perform our own GWAS using the ‘All of us’ Biobank to meta-analyse  or increase our GWAS collection for immunological, neurological infectous and cardiovascular diseases. We will then identify shared molecular mechanisms with Bayesian colocalisation analysis across diseases which outputs the probability of shared causal variants. The shared variants will be linked with publicly available expression QTLs, single-cell eQTLs (GTEx, Onek1k) and proteomics data (UK Biobank, PURE) to identify shared  molecular mechanisms and potential drugs for repurposing. Finally we will perform Polygenic Risk Scores (PRS) for co-morbidities using disease similarities and machine learning algorithms.

Aims

• Identify shared  biological mechanisms between diseases
• Find gene target candidates for drug repurposing and novel therapies
• Stratify population with a disease by risk of co-morbidities

Training Outcomes

• Analysis of large scale genomics and genetics datasets
• Software development skills
• Performing genetic associations in Biobank scale datasets
• Analysing eQTL, sceQTL and pQTL with Biobank scale datasets
• Integration of multiomics and genetic data
• Developing pipelines for functional genomics analysis to identify drug targets
• Application of statistical and machine learning models for drug targeting and patient stratification

Bibliography

1. DiMasi, J. A., Grabowski, H. G. & Hansen, R. W. Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal of Health Economics 47, 20–33 (2016).
2. Gallagher, M.D., Chen-Plotkin, A. The Post-GWAS era: From Association to Function. AHJG 102 pages717-730 (2018)
3. Pairo-castineira E. et al. Genetic mechanisms of critical illness in Covid-19. Nature 591, pages92–98 (2021)