Assessment of the CASP4 fold recognition category

Area of Science:

• Structural bioinformatics
• Computational biology
• Protein structure prediction

Background:

• The Critical Assessment of protein Structure Prediction (CASP) is a community-wide experiment to determine the accuracy of protein structure prediction methods.
• Fold recognition is a key challenge in structural bioinformatics, aiming to identify the correct three-dimensional structure of a protein from its amino acid sequence.

Purpose of the Study:

• To assess the performance of various computational methods in the CASP4 fold recognition category.
• To evaluate the accuracy and quality of predicted protein models compared to experimentally determined structures.

Main Methods:

• Multidomain targets were split into single domains for analysis.
• Predictions were classified, numerically evaluated, and mapped to quality indices.
• Performance was ranked using total and quality scores, considering both human groups and automated servers.
• The state-of-the-art in fold recognition was critically discussed.

Main Results:

• Most top-performing groups achieved high scores in both overall performance and model quality.
• Several predictors generated models significantly closer to target structures than known Protein Data Bank folds.
• The highest-ranked automated server achieved 12th place in the total score.
• Approximately two-thirds of all predictors demonstrated relatively low scores.

Conclusions:

• The CASP4 experiment highlighted the progress and capabilities of protein fold recognition methods.
• Top-performing groups demonstrated the ability to predict protein structures with high accuracy.
• Automated servers show promise, but further development is needed to match top human groups.