Bayesian statistical analysis of protein side-chain rotamer preferences

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Protein side-chain conformations are crucial for structure and function.
• Existing rotamer libraries have limitations in accuracy and scope.
• Backbone-dependent rotamer analysis is essential for precise modeling.

Purpose of the Study:

• To develop a statistically rigorous method for analyzing protein side-chain conformations.
• To create an extended backbone-dependent rotamer library.
• To validate the library using molecular mechanics and experimental data.

Main Methods:

• Bayesian statistical analysis of protein side-chain rotamer populations and chi angles.
• Incorporation of prior distributions based on previous data or pooled data.
• Modeling backbone-dependent rotamer probabilities using phi and psi angles.
• Validation using CHARMM22 molecular mechanics potential.

Main Results:

• A comprehensive rotamer library detailing populations and average chi angles for all phi, psi values.
• Demonstrated strong similarity between Bayesian analysis predictions and experimental distributions.
• Evidence that proteins favor lowest energy rotamers based on local backbone-side-chain interactions.

Conclusions:

• The developed Bayesian approach provides a robust statistical framework for rotamer analysis.
• The new rotamer library is a valuable resource for computational protein studies.
• The library supports applications in homology modeling, protein folding simulations, and structure refinement.