GIPPA

GIPPA (genome integrated prokaryotic proteomics analysis) takes one assembled bacterial genome plus proteomics evidence and asks: what does genome context explain about which proteins are detected, and how abundant they are? It places peptides and proteins on the genome with its own error control, builds one gene table keyed by Bakta locus tags, tests genome context (regions, autocorrelation, abundance islands, operons, function, sequence determinants), calls REF/ALT residues and novel ORFs, and writes standard tables, plots and an HTML report.

Documentation: khoa-yelo.github.io/gippa (from the first release; until then the same pages are under docs/). Changes between releases are listed in CHANGELOG.md. The process-level diagram shows every Nextflow process with its inputs and outputs.

GIPPA at a glance: from one genome and a peptide table to a gene table, genome-context tests and an HTML report, each claim with its own error control.

Status

1.0.0 is the first release. The full pipeline is tested on Linux x86-64 with Slurm and Apptainer, on the synthetic demo below and on the E. coli Nissle 1917 pilot (PXD078459; validation/pilot/README.md). Pending:

• the hosted docs and Bioconda;
• the final pilot report.

Genetic codes 11 (default), 4 (Mycoplasma and Spiroplasma) and 25 (SR1 and Gracilibacteria) are supported through annotation.translation_table (Inputs › Genetic code).

gippa preflight, convert, compare and report need only the Python package. gippa run also needs Nextflow, Java 17 or newer, Docker or Apptainer, the tool images and the databases (Install, Platforms).

Install

With pip (Bioconda is pending):

pip install gippa                                     # Python package and the `gippa` command
micromamba install -c conda-forge -c bioconda gippa   # from Bioconda, once the recipe is merged

From a checkout (Python >= 3.11):

git clone https://github.com/khoa-yelo/gippa.git && cd gippa
python3 -m venv .venv && . .venv/bin/activate && pip install .

gippa run also needs Nextflow (with Java 17+) and Docker or Apptainer; see Install for the prerequisites, the execution profiles and running on a Slurm cluster.

Quickstart

pip install gippa                         # or: conda install -c bioconda gippa
gippa demo                                # about 2 minutes; prints the report path
gippa setup                               # databases and images for the detected runtime
gippa run --genome g.fna --peptides p.tsv --outdir out

gippa demo runs a small bundled synthetic genome with no Nextflow, containers, databases or network. gippa setup and gippa run detect the Nextflow profile (pass --profile to choose). The guided walk-through, with example inputs and outputs, is Quick start (source). With raw spectra and no peptide table yet, see From raw spectra to a GIPPA run; for converting search-engine output, gippa convert.

Citing GIPPA

If you use GIPPA, please cite it as described in CITATION.cff (GitHub's "Cite this repository" button reads the same file).

Developers

The build specification is GIPPA_SPEC.md; progress is tracked in PROGRESS.md. Every PR that changes src/ or workflow/ adds a fragment under changelog.d/ (checked in CI by scripts/ci/check_changelog.py).

From a checkout, with uv:

uv sync --frozen --all-extras
uv run pytest -m "not slow and not calibration_full and not e2e" -n auto

See AGENTS.md for test layers and conventions.

Development on FarmShare

The maintainers' own cluster setup (scripts/farmshare_env.sh) is in the developer guide; users do not need it.

Licence

MIT. Third-party tools keep their own licences and run only as separate programs; see NOTICE.md.

geNOMAD (module mobile) is for non-commercial use only (gippa setup --no-genomad, --modules core,bgc,function); KEGG is free for academic use only (gippa setup --no-kegg); see Licences.