Late-stage drug functionalization is an approach to speed up the tedious process of pharmaceutical drug development, by starting from known drug templates and introducing or substituting chemical groups. In this project, the researchers aim to facilitate the challenging task of recognizing possible modification sites on template molecules with the help of machine learning.

The methodology is developed to predict better drug candidates that can activate target proteins associated with dysfunction of endothelial cells, which are the cells that form the inner lining of arteries and blood vessels. Endothelial cells dysfunction is a key factor in a range of vascular diseases for which currently no specific medical treatment exists.  

The synthesizability predictions by the developed AI algorithms are integrated with experimental validation in the organic chemistry laboratory. The biological activity of the predicted drug molecules will be assessed using blood-vessel-on-a-chip measurements in the UAMC hospital.

Researchers