ProPOSE: Direct Exhaustive Protein-Protein Docking with Side Chain Flexibility.

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Protein Interaction Analysis

Background:

• Protein-protein docking is crucial for understanding biological processes but remains challenging due to vast conformational search spaces.
• Existing methods like Fast Fourier Transform (FFT) approximate force fields, limiting accuracy.
• Predicting conformational changes and side chain rearrangements during binding is a key difficulty.

Purpose of the Study:

• To develop a direct search alternative for protein-protein docking that operates in real space.
• To improve the handling of side chain rearrangements during protein complex formation.
• To enhance the accuracy and native pose recovery in protein docking predictions.

Main Methods:

• Developed a direct search algorithm operating in Cartesian space, maintaining resolution and computational cost similar to FFT methods.
• Implemented strategies for focused exploration of conformational flexibility.
• Scored initial rigid poses while allowing for side chain overlaps, resolving them through subsequent repacking and minimization.

Main Results:

• The new method, ProPOSE, achieved significantly better native pose recovery on test systems.
• ProPOSE provided higher quality predictions (lower Root-Mean-Square Deviation - RMSD) compared to five leading docking programs.
• The approach effectively addressed challenges posed by side chain alterations in bound backbones.

Conclusions:

• Direct real-space search offers a viable and accurate alternative to FFT-based methods for protein-protein docking.
• ProPOSE demonstrates superior performance in predicting protein complex structures, particularly when side chain conformations change.
• The software ProPOSE (Protein Pose Optimization by Systematic Enumeration) is a promising tool for structural bioinformatics.