MolShaCS: a free and open source tool for ligand similarity identification based on Gaussian descriptors

Abstract

Molecular similarity evaluation is an important step in most drug development strategies, since molecular similarity is usually related to functional similarity. Here, we developed a method based on the Gaussian description of molecular shape and charge distribution for molecular similarity identification. The method was evaluated using the Directory of Useful Decoys (DUD) and a retrospective test. Enrichment factors computed for DUD targets showed that the proposed method performs very well in recognizing molecules with similar physicochemical properties and dissimilar topologies, reaching an average AUC of 0.63 and enrichment factor of 10 at 0.5% of decoys. A retrospective test also showed that nine mineralocorticoid ligands were ranked among the top ten molecules in a search of a database of approved drugs for molecules similar to aldosterone. Altogether, these data show that the Gaussianbased description of molecular shape and charge distribution implemented in the program MolShaCS is an efficient method for molecular similarity identification. The program is publicly available at the address http://www.ifsc.usp.br/biotechmol.