Template-Guided Protein Structure Prediction and Refinement Using Optimized Folding Landscape Force Fields.

Area of Science:

• Computational Biology
• Structural Biology
• Biophysics

Background:

• Template-based modeling is a reliable protein tertiary structure prediction method when structural templates are available.
• The Associative memory, Water mediated, Structure and Energy Model (AWSEM) is a coarse-grained force field optimized using energy landscape theory principles.

Purpose of the Study:

• To develop and validate a novel protein structure prediction method, AWSEM-Template, by combining template-based modeling with the AWSEM force field.
• To assess the accuracy and effectiveness of AWSEM-Template, particularly for targets in the low sequence identity "twilight zone" of homology modeling.

Main Methods:

• Incorporation of template information into the AWSEM force field using soft collective biases to create AWSEM-Template.
• Structure prediction tests on eight targets, including four in the twilight zone.
• Refinement of predicted structures using all-atom molecular dynamics with soft collective biases.

Main Results:

• AWSEM-Template achieved high-resolution protein structure prediction for all tested targets.
• The combination of AWSEM and template-guided potential improved prediction accuracy compared to template-guided potential alone.
• All-atom refinement further enhanced backbone and side-chain accuracy, reducing prediction errors and frustration.

Conclusions:

• AWSEM-Template effectively predicts protein tertiary structures, demonstrating superior accuracy especially when combined with AWSEM.
• The method's ability to maintain flexibility while guiding towards native-like structures is crucial for accurate predictions.
• All-atom refinement significantly improves the quality of predicted structures, enabling selection of highly accurate models.