Advances in AI-driven drug discovery through scaling cellular perturbation studies, coupled with machine learning, have accelerated the identification of potential therapeutic targets through use of in silico and in vitro approaches. However, many potential targets fail in clinical development due to poor efficacy and safety. This project will bridge this translational gap using UK Biobank’s large-scale multimodal data by validating AI-driven targets and prioritizing those with highest potential for clinical success.
We will achieve this goal through two aims:
Aim 1: Genetic Validation of Target Mechanisms. We will identify genetic variants associated with disease-relevant biomarkers and clinical phenotypes. Establishing these links will provide important evidence in patients to support or refute the proposed mechanism of action (MoA) for our novel targets.
Aim 2: Characterize Unmet Need and Patient Populations. We will conduct epidemiological analyses to benchmark outcomes on current therapies and characterize disease heterogeneity. This will define patient subgroups with the greatest unmet medical need, ensuring future clinical development is targeted effectively (e.g., by quantifying survival rates in specific cancer treatment settings).
This project will leverage the breadth and depth of UK Biobank, providing patient-centric evidence. This will enable us to prioritize the most promising therapies for advancement, directly addressing significant unmet medical needs.