Connecting the protein structure universe by using sparse recurring fragments

Area of Science:

• Structural biology
• Bioinformatics
• Computational chemistry

Background:

• Current protein structure classification relies on hierarchical schemes, often depicting protein fold space as discrete 'islands'.
• These schemes primarily focus on structural domains, potentially overlooking broader connections between protein folds.
• Recognizing shared structural and functional commonalities is crucial for understanding protein evolution and function.

Purpose of the Study:

• To propose an alternative perspective on protein fold space by introducing a novel similarity measure.
• To investigate the interconnectedness of protein structures based on shared fragments.
• To correlate inter-fold similarity with functional similarities across protein populations.

Main Methods:

• Development of an inter-fold similarity measure based on the frequency of shared sequence/structure fragments.
• Analysis of protein structure space using this novel similarity metric.
• Comparison of functional similarities between protein populations in different folds against the derived structural similarity.

Main Results:

• Protein fold space is visualized as a complex, interconnected network rather than isolated entities.
• The proposed inter-fold similarity measure effectively captures relationships between protein folds.
• Structural similarity, based on shared fragments, shows a strong correlation with functional similarity between protein groups.

Conclusions:

• The traditional view of isolated protein folds is an oversimplification; a network model is more accurate.
• Inter-fold similarity, derived from fragment sharing, provides valuable insights into protein structure-function relationships.
• This network perspective enhances our understanding of protein evolution and the organization of the proteome.