BioLance
Columnar variant store for multi-sample genomics.
BioLance ingests sample VCFs (and annotation VCFs like ClinVar) into a LanceDB directory where every field is an Arrow column. Query, join, and compare variants across samples in a single process — no Spark, no cloud, no GenomicsDB.
Docs
CLI reference, ingestion guide, and query examples.
Source
Apache-2.0 on GitHub.
What does BioLance do?
Standard bioinformatics pipelines produce one VCF per sample. Comparing variants across dozens or hundreds of samples means joining files that were never designed for cross-sample queries. BioLance solves this by converting VCFs into a Lance columnar store — one Arrow row per (sample, chrom, pos, ref, alt) — so region queries, ClinVar annotation joins, and multi-sample comparisons run out of a single directory with no infrastructure overhead.
Key commands: ingest,
query,
join,
compare,
screen,
compound-het, and
export.
Validated against real DeepVariant WGS data (GRCh37).