Identification of potential AChE inhibitors through combined machine-learning and structure-based design approaches

Abstract

Alzheimer’s disease (AD) is an irreversible, progressive neurodegenerative disease characterised by dementia.The depletion of acetylcholine (ACh) is involved the synaptic cleft is responsible for dementia due to neuronal loss. The acetylcholinesterase (AChE) enzyme isinvolved in the hydrolytic degradation of ACh and its inhibition is therapeutically beneficial for the treatment in memory loss.The use of machine learning (ML) for the identification of enzyme inhibitors has recently become popular. It identifies important patterns in the reported inhibitors to predict the new molecules. Hence, in this study, a set of support vector classifier-based ML models were developed,validated and employed to predict AChE inhibitors. Further, 247 predicted compounds obtained through PAINS and molecular property filters were docked on the AChE enzyme. The docking study identified compounds AAM132011183, ART21232619 and LMG16204648 as AChE inhibitors with suitable ADME properties. The selected compounds produced stable interactions with enzymes in molecular dynamics studies. The novel inhibitors obtained from the study may be proposed as active leads for AChE inhibition.