Leveraging Large-Scale Biobanks for Therapeutic Target Discovery, Predictive Ranking, and Drug Repurposing Opportunities

Large biobanks, including the Million Veteran Program (MVP), the UK Biobank, and FinnGen, provide genetic association results for more than 1,000,000 individuals for hundreds of phenotypes. To select targets for pharmaceutical development, as well as to improve the understanding of existing targets, we harmonized these studies, and performed two-sample Mendelian Randomization (MR) on a diverse set of phenotypes using genetic variants associated with gene expression and plasma protein levels as proxies of target modulation. We found 69,669 gene-trait pairs with evidence (p ≤ 1.6 x 10-9) for causal effects. From the selected gene-trait pairs, we observed 6,447 genes with strong evidence for at least one of 2,003 investigated traits. To enhance the translational potential of the resource, we developed a predictive ranking model trained using the drug targets of approved drugs described in ChEMBL 34 as well as several different biological annotations. This model was able to accurately predict the odds of a particular significant MR result being developed into an approved drug and its clinical indication (AUC 0.97, precision-recall AUC 0.79). Among selected results, we were able to rediscover 9% of approved drug targets in ChEMBL 34 establishing their relevance across different phenotypes. Moreover, our analysis revealed 900 drug-trait repurposing opportunities, where the identified genes were implicated for new therapeutic indications. Our model highlights the potential to modulate a broad range of phenotypes by targeting genes with high predicted rankings. By leveraging the integration of e- and p- QTLs with large-scale genetic studies and various biological annotations, we present a valuable approach for accelerating therapeutic development and expanding treatment options.