Discovering patterns and subfamilies in biosequences

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Biological sequence analysis often requires identifying conserved patterns and grouping related sequences into subfamilies.
• Existing methods may struggle with unaligned sequences or noise, limiting their effectiveness.
• The PROSITE database provides a valuable resource for protein family patterns.

Purpose of the Study:

• To develop a method for the simultaneous automatic discovery of patterns and subfamilies in unaligned biosequences.
• To address challenges posed by noisy and unaligned sequence data.
• To create a theoretically grounded significance measure for discovered patterns.

Main Methods:

• Utilizing the minimum description length (MDL) principle for pattern and subfamily discovery.
• Developing an algorithm to approximate the optimal set of patterns and corresponding subfamilies.
• Applying the method to unaligned biosequences with unknown noise levels.

Main Results:

• Successfully identified subfamilies within the chromo domain protein family.
• Discovered novel and significant patterns within the analyzed biosequence dataset.
• Demonstrated the effectiveness of the MDL-based approach in handling noisy sequence data.

Conclusions:

• The developed method provides a robust framework for simultaneous pattern and subfamily discovery in bioinformatics.
• The MDL principle offers a theoretically sound basis for evaluating pattern significance.
• This approach enhances the ability to analyze and classify protein families, aiding in functional annotation.