TOPAZ: asymmetric suffix array neighbourhood search for massive protein databases

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Protein homology search is crucial for annotation and metagenomics but challenged by growing databases.
• Existing methods like BLAST offer speed or sensitivity, but not both optimally.
• SANSparallel provides speed via parallel suffix array neighborhood search (SANS), with adjustable sensitivity.

Purpose of the Study:

• To develop a novel, highly sensitive, and rapid protein homology search method.
• To improve upon existing SANSparallel techniques for scalability and performance.
• To address the speed-sensitivity trade-off in large-scale biological database searches.

Main Methods:

• Introduction of asymmetric SANS for enhanced search capabilities.
• Implementation of scored seeds to improve detection accuracy.
• Development of an optimal substitution ordering scheme for suffix arrays.
• Benchmarking against BLAST using the UniProtKB database and a specific proteome.

Main Results:

• TOPAZ demonstrates substantial improvements in both sensitivity and speed over SANSparallel.
• The method achieves high performance in terms of speed, sensitivity, and scalability.
• TOPAZ searched UniProtKB for homologous proteins in under 3 minutes with 0.84 sensitivity, outperforming BLAST.

Conclusions:

• TOPAZ represents a significant advancement in protein homology search technology.
• It offers a superior balance between sensitivity and speed compared to existing methods.
• The tool is highly effective for large-scale homology detection in bioinformatics.