A fast protein-protein docking algorithm using series expansion in terms of spherical basis functions

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Biophysics

Background:

• Protein-protein interactions are crucial for biological processes.
• Accurate prediction of protein complex structures is essential for understanding function.
• Existing docking algorithms face challenges in speed and accuracy for unbound systems.

Purpose of the Study:

• To develop a fast and accurate ab initio protein-protein docking algorithm.
• To efficiently search the six-dimensional conformational space for rigid body molecules.
• To identify candidate conformations for unbound protein docking.

Main Methods:

• A novel series expansion using designed bases for conformational space search.
• A scoring function based on inner products of molecular scalar fields.
• Approximation of desolvation and steric hindrance energies using scalar fields.
• Fast Fourier Transform (FFT)-compatible scoring for flexibility in energy term incorporation.
• Basis function expansion (spherical harmonics and modified Legendre polynomials) for efficient scoring.

Main Results:

• The algorithm achieved high speed, with computations completed in approximately 40 seconds on a single CPU.
• Near-native conformations (interface RMSD < 3.0 Å) were identified within the top 1000 candidates for all six tested unbound cases.
• The method demonstrates precision comparable to existing approaches while significantly improving speed.

Conclusions:

• The developed algorithm offers a computationally efficient solution for ab initio protein-protein docking.
• The novel scoring and search strategy effectively explores conformational space for unbound complexes.
• This method holds potential for accelerating the discovery of protein interaction networks and drug design.