<![CDATA[Alzheimer’s disease is a progressive neurodegenerative disorder characterized by cognitive decline due to the accumulation of amyloid-β plaques in the brain. Beta-site amyloid precursor protein cleaving enzyme 1 (BACE1) plays a crucial role in forming amyloid-β peptides, making it a key therapeutic target in Alzheimer’s treatment development. This study aims to identify potential BACE1 inhibitors using a machine learning approach with the Random Forest and Extreme Gradient Boosting (XGBoost) algorithms, two distinct ensemble strategies (bagging and boosting). The dataset consisted of molecular representations as Morgan fingerprints (radius = 2, nBits = 2048), derived from compounds obtained through virtual screening and public databases such as ChEMBL. The models were trained and evaluated using a stratified 70:30 train–test split to maintain class balance across bioactivity levels (active, intermediate, and inactive). Model performance was evaluated using accuracy, F1-macro, precision-macro, recall-macro, and ROC-AUC-macro metrics. The results indicate that both models show strong classification capability in distinguishing BACE1 inhibitor activity, with XGBoost slightly outperforming Random Forest. Three compounds, CHEMBL3670318, CHEMBL5984885, and CHEMBL4740498, were predicted to be active and are potential candidates for further validation through molecular dynamics, in vitro, and in vivo studies. These findings suggest that ensemble-based machine learning can serve as an early-stage screening tool to prioritize BACE1 inhibitor candidates that still require experimental validation.]]>
Copyrights © 2026
