Cooperative approach for the protein fold recognition

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Protein Science

Background:

• Accurate protein structure prediction is crucial for understanding biological function.
• Identifying homologous or analogous folds for proteins with low sequence similarity remains a challenge.

Purpose of the Study:

• To assess the effectiveness of fold recognition methods in blind protein structure prediction.
• To evaluate the performance of a team-based approach combining multiple prediction tools and biological insights.

Main Methods:

• Utilized fold recognition techniques to predict structures for target sequences lacking obvious sequence homology.
• Employed three custom-developed threading programs with distinct compatibility functions.
• Integrated results from individual analyses with biological knowledge through collaborative discussions.

Main Results:

• Submitted 56 models for 25 target sequences, including novel fold predictions.
• Achieved successful prediction for 8 out of 18 threading targets (20 domains) at CASP3.
• Observed that combining predictions from multiple tools compensated for individual inaccuracies.

Conclusions:

• A collaborative, multi-method approach enhances protein structure prediction accuracy.
• Fold recognition is a viable strategy for predicting structures of proteins with distant evolutionary relationships.
• Effective management of diverse data sources is key to improving prediction reliability.