Searching databases of conserved sequence regions by aligning protein multiple-alignments

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Evolution

Background:

• Detecting evolutionary relationships between protein families is crucial for understanding protein function and evolution.
• Conventional sequence database searches have limitations in identifying weak or distant sequence relationships.

Purpose of the Study:

• To develop a general method for comparing multiple sequence alignments to detect sequence relationships between conserved protein regions.
• To identify previously undetected evolutionary links and functional insights within and between protein families.

Main Methods:

• Multiple sequence alignments are represented as sequences of amino acid distributions.
• Alignment is achieved by comparing pairs of these distributions using various comparison measures.
• The Pearson correlation coefficient was selected as the optimal comparison measure due to its sensitivity.

Main Results:

• The method successfully detected weak sequence relationships between diverse protein families, extending beyond the capabilities of conventional database searches.
• Identified a novel relationship between flavoprotein subunits of two oxidoreductase families, suggesting a potential active site.
• Revealed similarities among bacterial RecA, DnaA, and Rad51 protein families, pinpointing a likely DNA-binding region in DnaA and Rad51.
• Demonstrated the similarity of diverse helix-turn-helix DNA-binding domains and identified conserved features in glycosylasparaginase and gamma-glutamyltransferase enzymes.

Conclusions:

• The developed method is sensitive and effective for uncovering evolutionary relationships and functional information in conserved protein regions.
• This approach offers a powerful tool for exploring protein family evolution and discovering novel protein functions.
• The method is accessible via a World Wide Web implementation, facilitating broader research applications.