Installation

SCARAB 1.0.0 is available from the Hallam Lab Anaconda channel and Quay repository. Use Mamba for a native installation, Docker for a container, or Apptainer for a portable SIF image.

SCARAB supports Linux. All installation routes provide the same scarab CLI and use the dependency versions recorded in this repository. The supported runtime is Python 3.10; the scientific dependency versions are kept consistent across Python, Mamba, Conda recipes, and the container.

1. Released Mamba/Conda package

Install the release from the Hallam Lab channel:

mamba create -n scarab --strict-channel-priority \
  -c conda-forge -c bioconda -c hallamlab scarab=1.0.0
mamba activate scarab
scarab info
scarab recruit --help

The Conda package declares the Python libraries and external executables, including minimap2, SAMtools, MetaBAT’s depth summarizer, BBTools and SPAdes for guided reassembly. No Git-based pip helper installation is required. Use the pinned version above for the published workflow. The release checklist describes package validation and publication.

Then follow the same demo test used by all installation routes.

2. Docker

Pull the versioned image:

docker pull quay.io/hallamlab/scarab:1.0.0
docker run --rm quay.io/hallamlab/scarab:1.0.0 scarab info

Run the Docker demo command with your input directory mounted. Use a version tag or digest when recording a reproducible analysis. The Docker image installs the actual Python package; it does not depend on a source-directory PYTHONPATH wrapper.

To build this checkout locally:

docker build -t scarab:local .
docker run --rm scarab:local scarab recruit --help

BASE_IMAGE can be set to a pinned micromamba image digest for a release build. The build context allows only package source and installation files, excluding developer research directories and manuscript material.

3. Apptainer

Convert the versioned Docker image once, then move the SIF to the cluster:

apptainer pull scarab.sif docker://quay.io/hallamlab/scarab:1.0.0
apptainer exec scarab.sif scarab info

Run the Apptainer demo command. The SIF already contains SCARAB and its dependencies; an additional runtime Mamba environment is unnecessary. Build/pull on a machine with internet, then copy the SIF to a machine without internet access.

For an unpublished locally built image:

docker save scarab:local -o scarab-image.tar
apptainer build scarab.sif docker-archive://scarab-image.tar

This uses the local image archive and does not require uploading it to Quay.

4. GitHub source installation

git clone https://github.com/hallamlab/SCARAB.git
cd SCARAB
mamba env create -f environment.yml
mamba activate scarab_cenv
python -m pip install --no-deps --no-build-isolation .
scarab info
scarab recruit --help

environment.yml supplies the pinned scientific Python packages and native tools. Pip installs SCARAB into that active environment. Installing with pip alone does not install external mapping and deduplication programs. scarab_env.yml is a generated compatibility copy of the same environment specification, not a different installation route.

Run the demo test, then proceed to your own inputs. Before updating a working checkout/environment, stop any runs that use it. Keep separate output directories for different scientific configurations.

Installation verification

The 1.0.0 publication workflow tests the built Conda package, Docker image and converted Apptainer SIF using the same tiny recruitment-and-guided-reassembly demo. After publication, separate jobs install from the public Anaconda channel and pull the public Quay tag without registry credentials, then run the demo again. This checks both artifact contents and public distribution.