Ligand-Based Drug Discovery Leveraging State-of-the-Art Machine Learning Methodologies Exemplified by Cdr1 Inhibitor Prediction
Published in Journal of Chemical Information and Modeling (JCIM), 2025
Recommended citation: The-Chuong Trinh, Pierre Falson, Viet-Khoa Tran-Nguyen, and Ahcène Boumendjel. Ligand-Based Drug Discovery Leveraging State-of-the-Art Machine Learning Methodologies Exemplified by Cdr1 Inhibitor Prediction. Journal of Chemical Information and Modeling 2025. DOI: 10.1021/acs.jcim.5c00374 https://pubs.acs.org/doi/abs/10.1021/acs.jcim.5c00374
Abstract: This paper stems from my Master’s project and presents a comprehensive ligand-based drug discovery pipeline leveraging state-of-the-art machine learning methodologies, exemplified by the prediction of inhibitors for Candida albicans drug resistance 1 (Cdr1). Overcoming the challenge of multidrug resistance in fungal infections necessitates the discovery of novel Cdr1 inhibitors. In this work, we explored advanced techniques including Multi-Instance Learning (MIL) 3D Graph Neural Networks (3D-GNN) alongside traditional machine learning methods to identify potent Cdr1 inhibitors. The methodology demonstrates how cutting-edge AI can be effectively applied to screen and discover novel therapeutics.
Keywords: Multi-Instance Graph Neural Network, 3D-GNN, Machine Learning, Cdr1 Inhibitor, Cheminformatics, Ligand-Based Drug Discovery.