a16z Podcast: The Genetics Of Drug Delivery

Russ Altman presents the Pharmacogenomics Knowledge Base (PharmGKB), a 16-year effort to map how human genetic variation affects drug response, using multi-scale data integration from molecular interactions to electronic health records. He demonstrates how combining FDA adverse event reports, EHRs, and mouse models revealed unlisted drug interactions, such as paroxetine and pravastatin causing hyperglycemia. The framework enables drug repurposing, side effect prediction, and precision prescribing based on genetic profiles.

You get a concrete, multi-scale data framework for predicting drug effects, interactions, and repurposing opportunities grounded in real-world evidence and genetics.

1. 01Genetic variation significantly impacts drug metabolism—e.g., 7% of Europeans can't activate codeine due to a liver enzyme deficiency.
2. 02Combining data across biological scales (molecular, cellular, organismal, population) increases confidence in detecting real drug effects and interactions.
3. 03Unapproved drug combinations like paroxetine and pravastatin can cause significant side effects (e.g., hyperglycemia) undetected in traditional trials.