Data-driven drug discovery: investigating the molecular mechanisms of safety and efficacy

Project Summary I am proposing a precision pharmacology and pharmacovigilance research program that couples observational data analysis with prospective laboratory experiments to advance drug safety and efficacy. Our ability to collect and store massive amounts of molecular, clinical, and behavioral data has the potential to fundamentally transform translational medicine. It is not difficult to imagine a world where our devices and doctors work together seamlessly to provide personalized guidance and treatment to maximize our health and longevity. And that, in turn, the data generated by these encounters be collected, organized, and analyzed by biomedical researchers to invent the next generation of interventions. However, there are significant challenges prohibiting meaningful progress toward this vision. I have identified four that I plan to address: (1) There is a dearth of pharmacological knowledge for many subpopulations, most notably minorities (non-Whites), women, and children; (2) Observational data, from what is captured by devices to what is collected in medical records, is of dubious validity and value; (3) There is a limited understanding of the molecular mechanisms of drug reactions and drug-drug interactions; (4) There is no clear method of meaningfully sharing patient data while preserving privacy. There is no single solution that will solve all of these challenges. Each will require a unique combination of data science, informatics, and experiments. In the previously funded project, we made significant advancements in the characterization of adverse drug reactions and drug-drug interactions, the molecular modeling of pharmacological pathways, and the application of statistical data mining to electronic health records. I accomplished this by leveraging distinct data sources against each other to focus attention on only those hypotheses that repeatedly replicate under a variety of conditions. I then validated those hypotheses experimentally using animal and cellular models. Challenges 2 and 3 are natural extensions of this previous work, where I will address how to use data for purposes other than what it was collected for (secondary use) and develop new systems models to explain the physiological effects of drug-gene and drug-drug interactions. Challenges 1 and 4 represent new avenues of research where I will address the challenges of pharmacological studies in diverse populations and the increasingly important issue of balancing openness and transparency in science with patients' rights to privacy. The challenges laid out above are significant and, likely, will not be solved in within five years. However, the pursuit of these challenges will generate new knowledge that has the potential to significantly improve drug design, advance precision medicine, and guide drug safety governance.