Related Experiment Video
Updated: Jul 2, 2025
10:29
Quantitative Structure-Activity Relationship, Activity Prediction, and Molecular Dynamics of Non-nucleotide Reverse Transcriptase Inhibitors
Published on: May 9, 2025
1.1K
Machine learning and molecular simulation-based protocols to identify novel potential inhibitors for reverse
Muhammad Shahab1, Guojun Zheng1, Yousef A Bin Jardan2
1State Key Laboratories of Chemical Resources Engineering, Beijing University of Chemical Technology, Beijing, PR China.
Journal of Biomolecular Structure & Dynamics
|February 21, 2024
Summary
Machine learning models identified potential drug candidates for acquired immunodeficiency syndrome (AIDS) by targeting reverse transcriptase. The random forest model showed 86% accuracy, leading to the selection of three promising compounds for further study.
Area of Science:
- Computational chemistry and drug discovery
- Machine learning applications in pharmacology
Background:
- Acquired immunodeficiency syndrome (AIDS) is a critical condition caused by the human immunodeficiency virus (HIV).
- Inhibiting reverse transcriptase activity is a key therapeutic strategy for managing AIDS.
Purpose of the Study:
- To develop and evaluate machine learning models for identifying novel reverse transcriptase inhibitors.
- To screen compounds for potential anti-HIV drug development.
Main Methods:
- Utilized machine learning algorithms including support vector machines (SVM), k-nearest neighbor (k-NN), random forest (RF), and Gaussian naive Bayes (GNB).
- Trained models on a dataset of 5,159 compounds from BindingDB, with 1,645 active and 3,514 inactive against reverse transcriptase.
- Validated models using tenfold cross-validation and applied the best-performing model (RF) to an external ZINC dataset.
Main Results:
- The random forest model achieved 86% accuracy in predicting compound activity against reverse transcriptase.
- Three compounds (ZINC1359750464, ZINC1435357562, ZINC1545719422) were selected based on Lipinski's Rule, docking scores, and interaction analysis.
- Molecular dynamics simulations and MM/GBSA indicated stable interactions for the selected compounds with the reverse transcriptase enzyme.
Conclusions:
- Machine learning, particularly the random forest algorithm, is effective in identifying potential HIV reverse transcriptase inhibitors.
- The identified compounds represent promising leads for the development of new AIDS therapeutics.
