Universal Architectural Concepts Underlying Protein Folding Patterns.

Area of Science:

• Structural Biology
• Bioinformatics
• Computational Biology

Background:

• Understanding the fundamental building blocks of protein architecture is crucial for deciphering protein function and evolution.
• Existing methods often struggle to comprehensively categorize the diverse range of protein substructures.

Purpose of the Study:

• To identify and define a comprehensive dictionary of protein substructures, termed 'concepts,' that represent the fundamental architectural units of proteins.
• To explore the utility of these concepts in understanding protein function, evolution, and prediction.

Main Methods:

• Employed information-theoretic inference on an unbiased subset of protein structures to identify recurring substructures.
• Developed a dictionary of 1,493 'concepts,' defined as topologically conserved assemblies of helices and strands.
• Dissected the Protein Data Bank (PDB) to inventory all concept instances and analyze their correlations.

Main Results:

• A dictionary of 1,493 protein 'concepts' was established, representing the 'basis set' of protein architecture.
• Protein structures can be fully decomposed into instances of these concepts.
• Identified correlations between concepts and protein functions like catalysis and binding, and sequence-structure relationships.

Conclusions:

• The identified 'concepts' provide a fundamental framework for understanding protein architecture.
• This conceptual dictionary has significant implications for rational drug design, *ab initio* protein structure prediction, and evolutionary studies.
• An interactive resource, Proçodic, is available for accessing and navigating the concept dictionary and associated data.