Deriving protein dynamical properties from weighted protein contact number

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Atomic mean-square displacement (B-factor) in proteins correlates with neighboring atom counts (protein contact number).
• The protein contact model offers a computationally efficient way to calculate B-factors without complex simulations.
• This model is valuable for analyzing atomic fluctuations in large proteins and high-throughput studies.

Purpose of the Study:

• To refine the protein contact model by incorporating weighted contact numbers.
• To investigate the model's ability to compute cross-correlations of atomic motion.
• To assess the model's performance against experimental B-factors and other computational methods.

Main Methods:

• Developed a weighted protein contact-number model where weights are the square of reciprocal distances between contacting atom pairs.
• Applied the model to a dataset of 972 high-resolution X-ray protein structures.
• Computed B-factors and atomic motion cross-correlations using the refined model.

Main Results:

• The weighted protein contact-number model achieved a mean correlation coefficient of 0.61 between computed and experimental B-factors.
• This performance surpasses the original contact-number model (0.51) and other existing methods.
• Computed correlation maps showed global similarity to those from normal mode analysis.

Conclusions:

• The refined weighted protein contact-number model provides a more accurate prediction of protein atomic fluctuations.
• The study highlights a strong link between protein dynamics and protein packing.
• This method offers a powerful computational tool for exploring protein structure-dynamics relationships.