Structure classification-based assessment of CASP3 predictions for the fold recognition targets

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 assess protein structure prediction methods.
• Assessing the accuracy of protein structure prediction is crucial for understanding protein function and evolution.

Purpose of the Study:

• To evaluate the accuracy of protein structure prediction methods in the fold prediction/recognition category using CASP3 data.
• To analyze the relationship between predicted and experimentally determined protein structures, particularly for targets with no sequence similarity to known structures.

Main Methods:

• Analysis of predicted 3D structures against experimental structures for 23 CASP3 targets.
• Numerical evaluation and classification of predicted models to identify characteristic structural features.
• Comparative assessment of predictions to determine the best models for each target, including those with novel folds.

Main Results:

• A significant portion of targets, initially thought to have novel sequences, exhibited structural similarity to known proteins.
• Eleven of the 23 targets showed distant homology to known structures, and others had substantial similarities.
• The evaluation process successfully identified correct predictions and assessed the performance of different prediction methods.

Conclusions:

• The study highlights the challenge in identifying truly novel protein folds, as many targets share structural similarities with existing ones.
• The CASP3 results provided insights into the capabilities of various protein structure prediction methods.
• Accurate fold recognition remains a critical aspect of computational biology and drug discovery.