Projects

A Semi-Supervised Approach for Predicting Drug-Drug Interactions by Incorporating Drug Metabolism and Chemical Structure Knowledge

Using emerging technology to better predict drug-drug interactions.

Executive Summary

Drug-Drug Interactions (DDIs) become an emerging source of mortality and morbidity, threaten public health and drive up healthcare cost around the world. Unfortunately, the study of DDIs is challenging due to the complexity of drug structure and structure-metabolism relationship. Moreover, human genetic and demographic variations, patient medical conditions and other confounding factors make it more difficult to recognize and explain DDIs. With the improved availability of drug databases, adverse events reports, and electronic health records, data-driven machine learning tools are assuming importance and provide an opportunity to predict DDIs during new drug development and post-marketing surveillance.

Team Members

Qais Hatim (team lead), FDA
Kendra Woodbury, FDA
Sanjay Sahoo, FDA
Oanh Dang, FDA

Milestones

October 2018: Project selected into the HHS Ignite Accelerator
March 2019: Time in Accelerator Began
June 2019: Time in Accelerator Ended