Dictionary of recurrent domains in protein structures

Area of Science:

• Structural Biology
• Bioinformatics
• Computational Biology

Background:

• The increasing number of experimentally determined protein structures necessitates efficient classification methods.
• Protein structure classification is best performed at the domain level, representing evolutionarily conserved functional units.
• Existing methods struggle with consistent domain definition across distantly related proteins.

Purpose of the Study:

• To develop an automated method for identifying protein domains from atomic coordinates.
• To establish a comprehensive classification of known protein structures based on identified domains.
• To create a resource, the Dali Domain Dictionary, for exploring protein structure families.

Main Methods:

• Automated domain identification using quantitative measures of compactness and recurrence.
• Recursive division of proteins into substructures based on compactness.
• Selection of optimal substructure size using recurrence criteria for statistical significance.
• Clustering of identified domains to define fold classes.

Main Results:

• The method successfully identifies consistent domain definitions, even between remote protein homologs.
• Analysis of 1,137 protein families and 6,500 structures yielded 594 distinct protein fold classes.
• The Dali Domain Dictionary provides a global structural classification and links sequence families to structures.

Conclusions:

• The automated domain identification method provides accurate and consistent structural classification.
• The Dali Domain Dictionary serves as a valuable resource for studying protein evolution and function.
• This classification facilitates strategies for mapping the complete set of natural protein structures.