EROS-DOCK: protein-protein docking using exhaustive branch-and-bound rotational search

Area of Science:

• Computational biology
• Structural bioinformatics
• Biophysics

Background:

• Protein-protein docking is crucial for predicting 3D complex structures.
• Existing Fast Fourier Transform (FFT) docking methods face limitations with grid sampling and precision loss.
• Accurate protein complex structure prediction remains a significant challenge in structural biology.

Purpose of the Study:

• To introduce EROS-DOCK, a novel rigid protein-protein docking algorithm.
• To address limitations of FFT-based docking approaches.
• To enhance the accuracy and efficiency of protein complex structure prediction.

Main Methods:

• EROS-DOCK employs exhaustive 3D rotational searches using quaternion representation.
• A 'branch-and-bound' strategy systematically subdivides rotational space for efficient pruning of steric clashes.
• The ATTRACT coarse-grained force field model is utilized for scoring non-clashing orientations.

Main Results:

• EROS-DOCK was evaluated on 173 Docking Benchmark complexes.
• Performance was compared against established algorithms ATTRACT and ZDOCK.
• EROS-DOCK demonstrated a higher yield of acceptable or medium quality solutions based on CAPRI criteria.

Conclusions:

• EROS-DOCK offers an effective alternative to existing protein docking algorithms.
• The algorithm shows improved performance in predicting protein complex structures.
• EROS-DOCK provides a valuable tool for structural biology research.