Impact of structure space continuity on protein fold classification

Area of Science:

• Structural biology
• Bioinformatics
• Computational biology

Background:

• Protein structure classification aids in understanding protein structure-function relationships and evolution.
• SCOP and CATH are standard classification systems, but their discrete fold assumption is debated.
• Evidence suggests a continuous protein fold space, challenging current classification methods.

Purpose of the Study:

• To quantitatively evaluate the impact of continuous fold space on protein fold classification.
• To develop a novel method for classifying protein domains considering structure similarity and heterogeneity.
• To assess the differences between the new classification and existing schemes like SCOP and CATH.

Main Methods:

• Developed a likelihood-based method for domain classification.
• Incorporated both query-fold structure similarities and within-fold structure heterogeneities.
• Classified domains into existing CATH or SCOP folds.

Main Results:

• The new classification method reclassifies 3.4-12% of domains compared to original SCOP and CATH classifications.
• Discrepancies depend on structure similarity scores and the classification scheme used.
• The significance of structure space continuity varies with specific biological applications.

Conclusions:

• Current fold classifications may be limited by the assumption of discrete folds.
• A continuous fold space model offers a more nuanced approach to protein classification.
• The utility of considering fold space continuity is context-dependent for biological research.