AChEI-EL:Prediction of Acetylcholinesterase Inhibitors Based on Ensemble Learning Model

ABSTRACT

With the outbreak of the COVID-19, people are eager to develop potential drugs for specific diseases through efficient technological means. Alzheimer's disease (AD) has become one of the top ten causes of death in the world and is a typical neurodegenerative disease. When acetylcholinesterase (AChE) is inhibited, it improves the transmission of cholinergic neurotransmitters in patients and restores cognitive function, so acetylcholinesterase inhibitors (AChEIs) are often considered by researchers as potential drugs for the treatment of AD. Machine learning algorithms and data mining techniques can accelerate drug development and reduce the cost of biological experiments, so it is of great significance to develop models that can accurately predict AChEIs. However, few studies have applied efficient and mature ensemble learning methods to the problem of predicting potential inhibitors of AChE. In this study, we constructed a dataset from a publicly available biological experiment database, and for the first time established an ensemble learning model based on CatBoost and XGBoost to predict potential AChEIs. We demonstrate the advantages of ensemble learning models in building AChEIs predictor based on imbalanced, heterogeneous data through a comprehensive evaluation. Afterwards, we also combined the best-performing models into a blending model AChEI-EL for case studies, and obtained the top-ranked potential inhibitors that have been shown to have the potential to inhibit the AChE. These results suggest that our method has a promising application in the field of AD. Finally, we developed a WEB online prediction platform based on the best model for the use and reference of researchers. © 2022 IEEE.