Combining docking with pharmacophore filtering for improved virtual screening.

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• Virtual screening identifies potential drug leads from chemical libraries.
• Molecular docking predicts ligand-protein interactions but often yields inaccurate binding affinity rankings.
• Current scoring functions struggle to differentiate true binders from false positives.

Purpose of the Study:

• To develop a robust method for reducing false positives in virtual screening.
• To enhance the accuracy of identifying high-affinity ligands.
• To integrate docking and pharmacophore-based approaches for improved virtual screening efficacy.

Main Methods:

• Utilized molecular docking for pose generation, disregarding initial scoring.
• Employed receptor-based pharmacophore searches as a post-docking filtering step.
• Validated the method on three targets: neuraminidase A, cyclin-dependent kinase 2, and protein kinase C C1 domain.

Main Results:

• The combined docking and pharmacophore filtering approach successfully reduced false positives in virtual screening.
• This method demonstrated improved accuracy in identifying potential drug candidates.
• Performance was evaluated across multiple protein targets, showing consistent benefits.

Conclusions:

• Pharmacophore filtering following docking significantly outperforms traditional docking with scoring methods.
• This integrated approach leverages the strengths of both docking and pharmacophore modeling.
• The developed method offers a more reliable strategy for virtual screening in drug discovery.