Using pharmacophore models to gain insight into structural binding and virtual screening: an application study with...

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• Pharmacophore models predict activity based on feature presence and arrangement.
• Steric effects, crucial for ligand binding, are often unaccounted for in traditional pharmacophore models.
• Excluding volumes penalizes molecules in regions not occupied by active ligands, improving model accuracy.

Purpose of the Study:

• To apply the HypoGenRefine algorithm for automated generation of excluded volumes in pharmacophore models.
• To evaluate the impact of incorporating excluded volumes on model selectivity and virtual screening enrichment.
• To illustrate ligand-protein binding interactions within allowed and disallowed steric regions.

Main Methods:

• Utilized the HypoGenRefine algorithm within Catalyst software.
• Automated generation of excluded volume features from ligand information.
• Applied the methodology to two case studies: Cyclin-Dependent Kinase 2 (CDK2) and human Dihydrofolate Reductase (DHFR).

Main Results:

• Demonstrated the automated inclusion of excluded volumes to pharmacophore models.
• Illustrated how ligands can bind within protein active sites considering steric constraints.
• Achieved a more selective pharmacophore model with reduced false positives.
• Observed a better enrichment rate in virtual screening experiments.

Conclusions:

• Automated generation of excluded volumes using HypoGenRefine refines pharmacophore models by incorporating steric information.
• This approach enhances the accuracy of activity prediction and improves virtual screening efficiency.
• The method provides a more robust tool for drug discovery by better representing ligand-protein interactions.