Toward space-optimal approximate colored indexes at scale

Zindex builds interleaved ZOR filter colored indexes for 564,401 RefSeq assembled genomes in under 22 hours, producing a 289.5 GB index with cache-friendly throughput.

• Why it matters: Petabase-scale sequence repositories need compact, high-throughput colored indexes, but exact colored indexes and AMQ-based approaches can incur high construction/storage costs or suffer from scattered memory accesses.
• What they did: The framework uses interleaved ZOR filters with guaranteed construction at only ~1% above the AMQ space minimum, plus color-preserving fingerprint merging and a sketch-based layout planner.
• The result: It scales to 2,763,297 AllTheBacteria genomes with a 406.3 GB index in four days and a 43.8 GB peak memory footprint, enabling diverse seeding and sampling schemes.

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