Related Experiment Video
Updated: Sep 10, 2025
10:29
Quantitative Structure-Activity Relationship, Activity Prediction, and Molecular Dynamics of Non-nucleotide Reverse Transcriptase Inhibitors
Published on: May 9, 2025
1.5K
Benchmarking the Structure-Based Virtual Screening Performance of Wild-Type and Resistant PfDHFR Using Docking and
Menna S Hany1, Nermin S Ahmed2, Frank M Boeckler3,4
1Pharmaceutical Chemistry Department, Faculty of Biotechnology, German International University, Cairo, Egypt.
Drug Design, Development and Therapy
|August 21, 2025
Summary
This study benchmarks docking tools for malaria drug discovery. Machine learning re-scoring significantly improves identifying drug candidates against resistant Plasmodium falciparum Dihydrofolate Reductase (PfDHFR).
Area of Science:
- Computational chemistry and drug discovery
- Parasitology and infectious diseases
Background:
- Malaria, caused by Plasmodium falciparum, poses a significant global health threat.
- The parasite's Dihydrofolate Reductase (PfDHFR) enzyme is a crucial drug target.
- Mutations in PfDHFR confer resistance to antifolate antimalarials like pyrimethamine.
Purpose of the Study:
- To benchmark structure-based virtual screening (SBVS) tools against wild-type (WT) and mutant (Q) PfDHFR.
- To evaluate the impact of machine learning scoring functions (ML SFs) on SBVS performance.
- To provide recommendations for enhancing SBVS in antimalarial drug discovery.
Main Methods:
- Evaluated three docking tools (AutoDock Vina, PLANTS, FRED) using the DEKOIS 2.0 benchmark.
- Assessed two ML SFs (CNN-Score, RF-Score-VS v2) for re-scoring docking outcomes.
- Analyzed screening performance using pROC-AUC, pROC-Chemotype plots, and EF 1%.
Main Results:
- PLANTS with CNN re-scoring yielded the best enrichment for WT PfDHFR (EF 1% = 28).
- FRED with CNN re-scoring showed superior enrichment for the Q variant (EF 1% = 31).
- ML re-scoring substantially improved AutoDock Vina's performance for both variants.
- Re-scoring effectively retrieved diverse, high-affinity binders at early enrichment stages.
Conclusions:
- CNN-Score consistently enhances SBVS performance for both WT and Q PfDHFR.
- ML re-scoring improves the identification of diverse and high-affinity binders.
- These findings support improved antimalarial drug discovery, particularly against resistant strains.
