Machine learning-based virtual screening and molecular modelling studies for identification of butyrylcholinesterase inhibitors as anti-Alzheimer’s agent

Abstract

Butyrylcholinesterase (BChE) is a hydrolase involved in the metabolism and detoxification of specific esters in the blood. It is also implicated in the progression of Alzheimer’s disease, a type of dementia. As the disease progresses, the level of BChE tends to increase, opting for a major role as an acetylcholine-degrading enzyme and surpassing the role of acetylcholinesterase. Hence, the development of BChE inhibitors could be beneficial for the latter stages of the disease. In the present study, machine learning (ML) models were developed and employed to identify new BChE inhibitors. Further, the identified molecules were subjected to molecular property filters. The filtered ligands were studied through molecular modelling techniques, viz. molecular docking and molecular dynamics (MD). Support vector machine-based ML models resulted in the identification of 3291 compounds that would have predicted IC50 values less than 200 nM. The docking study showed that compounds ART13069594, ART17350769 and LEG19710163 have mean binding energies of −9.62, −9.26 and −8.93 kcal/mol, respectively. The MD study displayed that all the selected ligands showed stable complexes with BChE. The trajectories of all the ligands were stable similar to the standard BChE inhibitors. Communicated by Ramaswamy H. Sarma