CG-COA: a coarse-grained cubic orientation approach for systematic protein-protein docking validation

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Biophysics

Background:

• Traditional protein-protein docking methods struggle with conformational flexibility, impacting binding specificity.
• Accurately modeling protein flexibility is crucial for understanding biological interactions and drug design.

Purpose of the Study:

• To develop and validate a novel Coarse-Grained Cubic Orientation Approach (CG-COA) for protein-protein docking.
• To assess CG-COA's performance on systems with varying degrees of conformational flexibility and compare it with existing methods.

Main Methods:

• Implemented a Coarse-Grained Cubic Orientation Approach (CG-COA) combining coarse-grained molecular dynamics and MM/PBSA free energy calculations.
• Systematically sampled six cubic orientations for local docking.
• Validated the methodology on six diverse protein complexes, including intrinsically disordered proteins and systems with large conformational changes.
• Benchmarked against ZDOCK and ClusPro.

Main Results:

• CG-COA achieved a robust overall AUROC of 0.94 and F-scores up to 1.00 for well-defined systems.
• Demonstrated significant advantages for flexible systems (3.7-5.1-fold improvement for 3QML) but was outperformed by conventional methods for rigid interfaces.
• Identified limitations for extreme cases like intrinsically disordered proteins or large conformational changes, with poor structural accuracy (RMSDs >13 Å).
• Successfully diagnosed inherently flexible systems by not converging to single conformations.

Conclusions:

• CG-COA effectively addresses limitations of conventional docking for systems with moderate conformational flexibility.
• The method's applicability boundaries are clearly defined, showing it is unsuitable for extensive structural rearrangements or intrinsically disordered regions.
• CG-COA offers valuable guidance for protein interaction studies by identifying native binding modes and diagnosing system flexibility.