Centre for Computational Biology

Welcome

ANVAR project 2003

Virtual screening of chemical compounds for drug discovery

Principal investigators

•  Jean-Philippe Vert (Ecole des Mines de Paris).

Presentation

Developping new drugs is very expensive (about 5-800 millons USD per drug), but most of the cost is due to high failure rate in clinical trials because of toxicity or bad pharmacocinetic of the chemicals. The goal of this project is to develop new algorithms and tools for the virtual screening of chemical compouds at an early stage of the drug discovery process (even for virtual molecules not synthesized yet), in order to reduce the failure rate in later stages. This involves new algorithms for predicting toxicity, activity and drugability of candidate molecules in silico.

The methods developed are based on recent advances in statistical learning theory and kernel methods.

Support

This project is supported by the French National Agency for Innovation.