2017 SOT Monday Daily Plenary Session: Data Science

Systems Approaches to Drug Efficacy and Toxicity in an Era of Big Data

Lecturer: Peter Sorger, Harvard Medical School, Boston, MA.

The development of new therapeutic drugs is fundamental to improving human health, but the process is challenged by rising costs and a high rate of failure. New and better technology and big data are often put forward as the solutions to these problems. However, I will discuss laboratory and clinical studies showing that some of the fundamental concepts in pharmacology and toxicology are ripe for reinvention. Increasing data on the impact of cell-to-cell variability and temporal variation in cellular physiology motivates new ways of thinking about seemingly simple concepts such as drug dose-response. Better understanding of sources of variation in laboratory and clinical data should also improve our ability to identify robust biomarkers of therapeutic and adverse effects.

I will argue that big data and data science are essential but insufficient: correct interpretation of empirical data in biomedicine hinges on theories about mechanism. I will discuss these theories, with reference to cytotoxic and targeted anti-cancer therapies, and studies of drug response in cell culture, animal models, and human clinical trials. New pharmacological principles derived from such studies are being developed into practical algorithms and open-source software as a means to improve target qualification, lead molecule optimization, and early phase clinical trials. The hoped for outcome: better drugs at a cost society can afford.

Open Ecosystems for Understanding Toxicities and Adverse Events

Lecturer: Lara Mangravite, Sage Bionetworks, Seattle, WA.

The presentation will address the use of collaborative approaches for the gathering, sharing and interpretation of health data. This will include the use of remote sensor-based data collection approaches to capture fluctuations in health relative to medication and disease—and the consideration of how these could be used to track adverse events.