This site is devoted to computer and chemoinformatics research carried out within the framework of a close collaboration between two academic laboratories :
- GREYC • computer laboratory, UMR CNRS 6072
- CERMN • medicinal chemistry laboratory, UNICAEN UR 4258, FR CNRS 3038 INC3M.
The collaboration focuses on the design of computational methods to identify chemical (sub)structures of interest (for activity, toxicity, …) in the early stage of drug discovery. For this purpose, we rely on bioactivity data stored in public chemical databases such as ChEMBL () or in-house data (chemical library of CERMN).we have information concerning the affinity of ligands for targets (a ligand is a molecule that binds to a target).
From this initial information, we calculate statistical associations highlighting chemical (sub)structures whose presence seems to influence the interaction of a molecule with the target(s) of interest. In addition to the classical representation of molecules, we work at the pharmacophore level and this is an important point for our latest developments.
The chemoinformatics methods we are developping allow 1) to automatically extract substructures or pharmacophores from a large dataset of molecules, 2) to integrate data mining techniques designed for the resulting descriptors. These methods are currently available through our software Norns which is useable with an open access license. Norns is a graphical environment dedicated to programming chemoinformatics workflows. In its current version, Norns contains many configurable components for elementary operations (pharmacophore extraction, performance measures calculation, visualization, …). Moreover, the environment has been designed to allow the user to easily develop their own boxes.