Database of bioactive ring systems with calculated properties and its use in bioisosteric design and scaffold hopping

Abstract

A system for identification of bioisosteric scaffolds is presented. The system uses a database of over 7000 scaffolds extracted from bioactive molecules. Scaffolds in the database are characterized by their size, shape, pharmacophore features and several ADME descriptors. Also properties characterizing electron-donating or -accepting power at connection vectors are considered. All these features are used as search criteria to find scaffolds with the most similar properties to the query. To guarantee fast processing the search is performed using topological descriptors only, but the system may be used to find optimal replacements of scaffolds also directly in the protein binding site. In this case a set of 3D conformations for the best 2D hits is generated and analogs optimally fitting the binding pocket are identified by overlap with the query ligand and by optimizing interactions with the protein. This tool is used at Novartis as an idea generator for identification of novel non-classical bioisosteric analogs in the drug discovery process.