Installation

Requirements

• Python ≥ 3.10
• Linux or macOS (pysam wheels are not published for native Windows — use WSL or conda there).
• An indexed reference FASTA (.fai alongside it; samtools faidx ref.fasta) — or internet access to use --ref-source api (UCSC) with no local FASTA.
• Optionally, gnomAD VCFs or internet access to the gnomAD API (for masking).

From PyPI

pip install vflank

From GitHub (latest, unreleased)

pip install git+https://github.com/rhshah/vFlank.git

From source (development)

git clone https://github.com/rhshah/vFlank.git
cd vFlank
pip install -e ".[dev]"

Runtime dependencies (typer, rich, pysam, pandas) install automatically.

Verify

vflank version
vflank --help

You should see the small and fusion command groups.

Reference and masking data

vflank never bundles large genomic data. You provide:

• a reference matching your --genome-build (hg19/GRCh37 or hg38/GRCh38): either a local indexed FASTA (--ref-source file, default) or the UCSC API (--ref-source api, no download);
• for masking, either a directory of gnomAD per-chromosome VCFs (--pop-source vcf) or nothing at all when using --pop-source api.

See SNP Masking for details and download commands.