Flexible protein docking refinement using pose-dependent normal mode analysis.

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Biophysics

Background:

• Protein docking is crucial for understanding molecular interactions but modeling conformational changes remains a challenge.
• Current methods often treat proteins as rigid or use computationally intensive ensemble approaches.
• Existing techniques struggle with the vast number of poses generated, requiring sensitive scoring functions.

Purpose of the Study:

• To introduce EigenHex, a novel computational method for protein docking that incorporates protein flexibility.
• To evaluate the effectiveness of pose-dependent normal mode analysis (NMA) in refining docking poses.
• To provide a more efficient alternative to combinatorial cross-docking methods.

Main Methods:

• Conventional rigid-body docking using the Hex algorithm.
• Application of pose-dependent NMA based on an elastic network model to refine top docking solutions.
• Sampling and re-scoring of perturbed conformations using linear combinations of eigenvectors and particle swarm optimization.

Main Results:

• EigenHex demonstrated a modest but consistent improvement in docking accuracy for rigid-body targets.
• Utilizing just one to three eigenvectors in NMA was sufficient to enhance performance.
• The method effectively avoids the computational burden of sampling numerous eigenvectors.

Conclusions:

• Pose-dependent NMA offers a computationally efficient way to model protein flexibility in docking.
• EigenHex presents a promising advancement over traditional rigid-body docking and combinatorial cross-docking.
• This approach enhances the accuracy of protein-protein interaction predictions.