Local Protein Structure Refinement via Molecular Dynamics Simulations with locPREFMD.

Area of Science:

• Structural biology
• Computational chemistry
• Biophysics

Background:

• High-resolution protein structures are crucial for understanding biological function.
• Local stereochemical inaccuracies can limit the utility of computational models.
• Maintaining the overall protein fold during refinement is essential.

Purpose of the Study:

• To present a novel method for the local refinement of protein structures.
• To improve local stereochemistry while preserving the global protein fold.
• To provide a user-friendly web server implementation for this method.

Main Methods:

• Utilized force field-based minimization and molecular dynamics simulations.
• Employed a modified force field for targeted refinement of bonds, angles, and torsion angles.
• Implemented the method in the locPREFMD web server and validated on CASP11 models.

Main Results:

• The locPREFMD method significantly improved stereochemical quality of protein models.
• Refined models approached the quality standards of experimental structures.
• The Cα coordinates of the initial models were successfully preserved.

Conclusions:

• The locPREFMD method offers an effective approach for local protein structure refinement.
• The method enhances stereochemical accuracy without compromising global structural integrity.
• The locPREFMD web server provides a valuable tool for computational structural biology.