Extracting and Updating Compact Colored k-mer Indexes

Updating spectral Burrows-Wheeler transform (SBWT) colored k-mer indexes in batches enables genome removal and extraction with rebuild-equivalent results, cutting E. coli index edits from 13 minutes to 54 seconds.

• Why it matters: Colored k-mer indexes are costly to rebuild after genome retractions, contamination fixes, or when users need a single species or clade, yet collections change over time.
• What they did: The method extracts the SBWT for any subset of k-mers in one linear pass from an existing SBWT, then propagates colors through removal, intersection, difference, and union using only the inputs.
• The result: In experiments, removing one genome from an index of 3,682 Escherichia coli genomes took 54 seconds versus 13 minutes for rebuild, adding one took about a minute, and extracting Mycobacterium tuberculosis from AllTheBacteria took under four hours.

AI summary of the abstract · may contain errors · not medical advice · read the paper.