Local multiple alignment of numerical sequences: detection of subtle motifs from protein sequences and structures

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Identifying conserved motifs in proteins is crucial for understanding protein function and evolution.
• Existing methods often rely solely on sequence data, potentially missing structurally conserved but sequence-divergent motifs.

Purpose of the Study:

• To develop and validate a new computational method for motif discovery integrating protein sequence and structure information.
• To improve the accuracy and scope of motif identification compared to sequence-only approaches.

Main Methods:

• A two-part method involving quantification of protein sequences and structures into numerical representations.
• Development of a novel Gibbs sampling algorithm adapted for real number/vector sequences to find common regions.
• Local multiple alignment to identify fixed-length regions with similar shapes.

Main Results:

• The new method successfully identifies motifs from both multiple protein sequences and structures.
• The integrated approach demonstrates particular utility when structural information is available.
• Comparison with a standard Gibbs sampling program highlights the advantages of the new method.

Conclusions:

• The proposed method offers a powerful new tool for motif discovery in bioinformatics.
• Integrating structural data significantly enhances the identification of conserved protein motifs.
• This approach has implications for functional annotation and drug discovery.