Development and application of ligand-based computational methods for de-novo drug design and virtual screening

Abstract

Ligand-based computational drug discovery (LB-CADD) methods have been used widely over the last several decades to aid medicinal chemistry campaigns via virtual high-throughput screening (vHTS) and de-novo molecular design. A new de-novo drug design algorithm, BCL::EvoGen, based on a stochastic search algorithm was implemented within the BioChemical Library developed at Vanderbilt University. The EvoGen algorithm leverages reaction-based structure modification methods to iteratively build chemical structures, and ligand-based molecule scoring functions to guide molecular design. Results indicate that the EvoGen algorithm is capable of designing high-scoring molecules with novel and chemically reasonable structures. In a second study, LB-CADD models were used to prioritize a subset of a compound library the discovery of muscarinic acetylcholine receptor M5 negative allosteric modulators. An orthosteric antagonist VU0549108 (VU108) was discovered which exhibited an M5 IC50 of 5.23 uM and moderate selectivity across other muscarinic receptors. In addition, VU108 contains a novel chemical scaffold not previously associated with muscarinic receptor ligands.