Div-BLAST: diversification of sequence search results.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Sequence similarity tools like BLAST are crucial for bioinformatics research.
• Current tools often return highly similar, redundant results, limiting exploratory analysis.
• There's a need for methods that introduce diversity into sequence similarity search results.

Purpose of the Study:

• To introduce and address the problem of diverse search and browsing in sequence databases.
• To define sequence diversity measures and propose methods for obtaining diverse results.
• To develop a new measure for evaluating the diversity of query results.

Main Methods:

• Proposed methods for obtaining diverse results by post-processing BLAST and PSI-BLAST.
• Developed a novel measure to evaluate the diversity of sequence sets.
• Assessed functional diversity using Gene Ontology annotations.
• Compared performance against CD-HIT, a redundancy elimination tool.

Main Results:

• The proposed methods generate more diverse and significant result sets compared to static non-redundancy approaches.
• Both sequence-based and functional diversity evaluations showed significant improvement over original BLAST and other baselines.
• The Div-BLAST web tool implements these diversification methods.

Conclusions:

• Incorporating diversity into sequence similarity searches significantly enhances bioinformatics exploration.
• The developed methods outperform existing approaches in generating diverse and functionally relevant results.
• Div-BLAST offers a valuable tool for researchers seeking broader insights from sequence similarity searches.