ont-bed-generator 0.4.0

Generate BED files for Oxford Nanopore adaptive sampling from a gene list and a RefSeq GFF3.

ont-bed-generator

Generate BED files for Oxford Nanopore adaptive sampling from a gene list and a RefSeq GFF3. Standalone reimplementation (Python standard library only, no runtime dependencies) of the Galaxy workflow "Adaptative ONT BED file generation" (which replaced BED-Craft).

Given a list of target genes and a GFF3 annotation, the tool produces:

• targets.bed — the raw locus of each gene (BED6);
• merged-extended.bed — two intervals per locus (+ / - strands), extended, clamped to chromosome bounds, then merged per strand.

Installation

# from source
git clone https://github.com/CHANGE-ME/ont-bed-generator.git
cd ont-bed-generator
pip install .

Conda development environment (includes bedtools, used only by the parity test):

conda env create -f environment.yml   # or: mamba/micromamba env create -f environment.yml
conda activate ont-bed-generator

Usage

ont-bed-generator \
    --genelist genelist.tsv \
    --gff annotation.genes.gff3 \
    --genome hs1.len \
    --out-targets targets.bed \
    --out-merged  merged-extended.bed
# or: python -m ont_bed_generator --genelist ... --gff ... --genome ...

Inputs

1. genelist TSV — columns Gene | Left_extension_bp | Right_extension_bp, with an (auto-detected) header row. Coordinates come from the GFF, so no chromosome column is needed; a gene is treated as an extended region when Left or Right is non-zero (no separate flag). A bare Gene line (no extensions) is valid and gets only the default flank. Symbols must be official HGNC symbols (check them with the multi-symbol checker).
2. GFF3 — a RefSeq GFF3 whose gene features carry Name= and Dbxref=GeneID:. The full NCBI annotation works as-is: the tool reads only gene features and ignores everything else, so pre-extracting them is just an optional speed-up. Gzip-compressed (.gz) input is read transparently (this applies to every input file). Example (T2T-CHM13v2.0 / hs1):

   curl -sO https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/009/914/755/GCF_009914755.1_T2T-CHM13v2.0/GCF_009914755.1_T2T-CHM13v2.0_genomic.gff.gz
   

3. genome sizes — chrom<TAB>size, one line per chromosome. Used for telomere clamping only; the line order does not matter (output BEDs are sorted in canonical karyotypic order).

   curl -sO https://hgdownload.soe.ucsc.edu/goldenPath/hs1/bigZips/hs1.chrom.sizes
   

Symbol resolution

Official symbol → GeneID (GFF Name=, or external --entrez-map) → collect all loci carrying that GeneID. The Entrez key handles pseudoautosomal regions (one GeneID on chrX and chrY yields two loci) and IG/TCR loci. A symbol that is not found, or that maps to several GeneIDs, is reported (on stderr), never guessed. --strict returns a non-zero exit code if any symbol is invalid or ambiguous.

Development

pip install -e '.[dev]'
pytest            # unit tests (+ bedtools parity if bedtools is present)
ruff check .
mypy

Versioning is driven by git tags (hatch-vcs): a release is git tag vX.Y.Z && git push --tags.

Intentional differences from the original Galaxy workflow

• The workflow patched non-official symbols (MKL1→MKLN1, a bug — MKLN1 is a different gene; LYL→LYL1). Here no patching: such symbols are reported for correction at the source.
• The Galaxy output contained a missing merge (bedtools merge on genome-order-sorted input without -g). This tool performs the correct merge.

License

MIT — see LICENSE.