CASP 11 target classification.

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein structure prediction

Background:

• The Critical Assessment of Structure Prediction (CASP) is a community-wide experiment to assess protein structure prediction methods.
• Accurate classification of protein targets is crucial for evaluating different modeling approaches.

Purpose of the Study:

• To classify protein target structures from CASP11 and CASP ROLL into categories for assessing template-based modeling (TBM) and free modeling (FM).
• To establish criteria for distinguishing between TBM and FM targets based on evolutionary relatedness and prediction performance.

Main Methods:

• Protein target structures were split into domain-based evaluation units.
• Classification into TBM and FM categories was performed using evolutionary relatedness (ECOD database) and server prediction performance.
• Criteria considered template detectability and sequence similarity to existing structures.

Main Results:

• Target structures were successfully divided into domains and classified into TBM and FM categories.
• The FM category included 45 target domains, an unprecedented number for CASP.
• Classification boundaries were sometimes blurred due to insertions, deletions, or partial template coverage.

Conclusions:

• The study established a robust classification scheme for CASP targets, differentiating between TBM and FM.
• The high number of FM targets highlights the increasing difficulty and importance of free modeling in protein structure prediction.