CASP9 results compared to those of previous CASP experiments

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Protein Modeling

Background:

• The Critical Assessment of protein Structure Prediction (CASP) experiments benchmark protein structure modeling methods.
• Previous CASP assessments have tracked progress in computational protein structure prediction.

Purpose of the Study:

• To analyze the quality of structure models submitted to CASP9.
• To evaluate progress in protein structure modeling compared to previous CASP rounds.
• To introduce and apply new methods for assessing model quality.

Main Methods:

• Analysis of CASP9 structure models using established comparison techniques.
• Application of new methods to assess model quality in regions without clear templates.
• Evaluation of alignment accuracy and progress in template-free modeling.

Main Results:

• Overall progress in CASP9 was modest and statistically challenging to validate.
• Improvements observed in mid-range template-based modeling difficulty.
• Enhanced methods for selecting the best model from generated sets.
• Significant progress indicated for template-free models of short proteins.
• Better performance identified in model regions not covered by the best template.

Conclusions:

• CASP9 demonstrated positive trends, particularly in template-free modeling and assessing model quality in challenging regions.
• While overall progress is modest, specific areas show promising advancements.
• Improved methods enhance the evaluation of protein structure models.