Surflex-Dock 2.1: robust performance from ligand energetic modeling, ring flexibility, and knowledge-based search.

Area of Science:

• Computational chemistry and cheminformatics
• Drug discovery and molecular modeling

Background:

• Traditional molecular docking methods often have limitations in accurately representing ligand flexibility and guiding search processes.
• Previous Surflex versions focused on alignment and torsional space, limiting the exploration of complex ligand conformations.

Purpose of the Study:

• To enhance the Surflex flexible molecular docking method by improving its search capabilities for increased accuracy and screening utility.
• To incorporate ligand energetics and knowledge-based interactions to guide the docking search process more effectively.

Main Methods:

• Incorporation of a small-molecule force field to extend the search into Cartesian coordinates, enabling dynamic ring flexibility and all-atom optimization.
• Integration of knowledge-based molecular interactions between ligand fragments and target proteins to guide the search process.
• Validation using multiple publicly available benchmark datasets to assess docking accuracy and screening utility against other methods.

Main Results:

• Surflex-Dock 2.1 demonstrates comparable docking accuracy to the best available molecular docking methods.
• The enhanced Surflex method shows robust and superior performance in screening utility, achieving approximately double the screening enrichment.
• The new approach supports dynamic ring flexibility and all-atom optimization of docked ligand poses.

Conclusions:

• The generalized Surflex method, with its extended search capabilities, offers significant improvements in both docking accuracy and screening utility.
• Surflex-Dock 2.1 provides a powerful and robust tool for drug discovery, outperforming other methods in screening enrichment.
• The integration of force fields and knowledge-based interactions represents a significant advancement in flexible molecular docking.