The abstract is read at the publisher; the summary is JClub's.
pbzarr stores per-base genomic signals for cohorts as a single Zarr store, achieving up to 2.1x faster cohort construction while using 5-6x less memory across multiple datasets.
- Why it matters: Cohort-scale per-base analyses are slowed by formats built for single samples, requiring opening and merging thousands of files before downstream computation.
- What they did: The toolkit converts many per-base signal tracks into native Zarr arrays, enabling cohort construction via a single command and supporting region queries and full-genome scans through array-based access.
- The result: For 1,024 short-read genomes, it stores per-base coverage using only 3.3% of the input size, with region-based queries 61x faster than D4 and 6x faster than bigWig.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.