Combining co-evolution and secondary structure prediction to improve fragment library generation

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Accurate protein structure prediction is crucial for understanding biological function.
• Template-free protein structure prediction methods have advanced significantly.
• Co-evolution techniques offer new avenues for improving prediction accuracy.

Purpose of the Study:

• To develop a general approach for generating improved fragment libraries for protein structure prediction.
• To utilize co-evolution constraints to enhance fragment library quality.
• To assess the impact of improved fragment libraries on protein structure modelling.

Main Methods:

• Comparison of five fragment library generation programs across three datasets (>400 protein folds).
• Incorporation of secondary structure information for fragment library assessment.
• Application of co-evolution constraints to enrich and refine fragment libraries.

Main Results:

• Fragment libraries considering secondary structure show improved usefulness for prediction.
• Co-evolution constraints successfully enriched libraries by selecting compliant fragments.
• Improved fragment libraries led to higher precision and consistently better protein modelling results.

Conclusions:

• Co-evolutionary constraints are effective in generating high-quality fragment libraries.
• The developed approach significantly enhances fragment-based protein structure prediction.
• The Flib-Coevo tool and data are publicly available for further research.