High-resolution structure prediction and the crystallographic phase problem.

Area of Science:

• Computational structural biology
• Protein structure prediction
• X-ray crystallography

Background:

• Refining low-resolution protein models to atomic accuracy is a significant challenge.
• Accurate protein structures are crucial for understanding biological function and drug discovery.

Purpose of the Study:

• To develop and validate a novel computational approach for refining protein structure models.
• To improve the accuracy of protein structures derived from various sources, including experimental and comparative modeling.
• To assess the utility of refined models in solving the X-ray crystallographic phase problem.

Main Methods:

• An energy-based refinement method focusing sampling on error-prone regions.
• Utilizing a physically realistic all-atom force field for structure relaxation.
• Application to models from nuclear magnetic resonance (NMR) data and comparative modeling.

Main Results:

• Significant improvement in backbone conformations and side-chain placement accuracy.
• Enhanced performance in solving the X-ray crystallographic phase problem via molecular replacement.
• Successful de novo protein structure prediction achieving molecular replacement accuracy without experimental phase information.

Conclusions:

• The developed refinement method substantially enhances protein model accuracy.
• This approach offers a powerful strategy for tackling challenging molecular replacement phasing problems.
• It holds promise for advancing de novo protein structure prediction and experimental structure determination.