AnGe is a tool for generating bioisosteric analogs of a user-submitted query molecule. The molecule is first decomposed into its constituent parts - such as scaffold, rings, functional groups, linkers, and substituents. For each of these components, a set of bioisosteric analogs is identified based on matched molecular pairs extracted from the ChEMBL database, and similarity in calculated molecular properties, including ADME characteristics and quantum chemically derived electronic parameters.

Compatible analogs are then constructed by recombining these substructure components. This generation process can be repeated iteratively, each cycle expanding the exploration of chemical space and yielding a broader set of analogs.

The web tool available on this page provides only a limited demonstration of AnGe capabilities. The full-featured Java engine allows much finer control of the generation process, including settings for the number of iterations, and constraints on substructure features to retain or exclude.

The core concept behind AnGe is to offer a large, diverse pool of chemically plausible analogs, which can subsequently be filtered or prioritized using additional criteria such as property filters or molecular docking.