Indexing and searching petabase-scale nucleotide resources.

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Vast nucleotide sequence data in resources like the Sequence Read Archive and GenBank are difficult to search.
• Current methods are impractical for most researchers to navigate large-scale genomic data.

Purpose of the Study:

• To introduce Pebblescout, a novel tool for indexing and searching large nucleotide sequence resources.
• To enable efficient navigation and retrieval of relevant data from massive genomic datasets.

Main Methods:

• Pebblescout employs dense sampling for indexing nucleotide sequences.
• Search functionality identifies relevant runs or assemblies based on short sequence matches.
• Results are ranked by the informativeness of the matches, providing well-defined guarantees.

Main Results:

• Eight databases indexing over 3.7 petabases were created using Pebblescout.
• The tool effectively finds relevant subsets of large nucleotide resources across various query lengths.
• Pebblescout demonstrates favorable performance compared to existing tools like MetaGraph and Sourmash.

Conclusions:

• Pebblescout significantly reduces the effort required for downstream analysis of large nucleotide resources.
• The tool offers a data-driven approach for efficient exploration of massive genomic data.
• Pebblescout provides a practical solution for researchers dealing with rapidly growing sequence databases.