Despite advancements in biomedical research, drug development success rates remain low, particularly for neurological, oncological, and rare diseases. These areas face significant unmet needs due to small, heterogeneous patient populations and complex disease mechanisms. Integrating human genetics, multi-omic data, and real-world clinical trajectories can improve outcomes by providing insights that increase the probability of success.
This project addresses these challenges by leveraging population-scale genetic and multi-omic data to: identify novel drug targets and biomarkers, define optimal patient populations and trial endpoints for clinical translation, and explore opportunities for drug repurposing: First, we apply clinical expertise, epidemiology, and advanced machine learning algorithms, including transformer-based models, to deeply characterize patient populations using clinical, multi-omic, imaging, and survey data. Second, genome- and exome-wide association studies will identify mechanisms underlying these defined subtypes, while phenome-wide association studies and Mendelian Randomization uncover causal relationships between targets, biomarkers, and outcomes. Third, we will generate hypotheses for interventions on specific subtypes and assess their expected impact using genetic evidence. We will investigate biomarkers (i.e. metabolites, protein-levels, and clinical markers) and combinatorial disease signatures derived from multi-omic data to stratify patient populations. Ultimately, we will investigate the robustness of our findings to enable discovery and clinical translation by considering data from past clinical studies and other real-world datasets using epidemiological and comparative analyses.
We expect this comprehensive approach to catalyze critical insights into disease biology, which directly impact the development of novel therapeutics – thereby immediately benefitting patients with high unmet needs in challenging disease areas.