Using pyloseq in Containers

pyloseq has no dedicated Docker image - it doesn't need one. Because it's a pure Python package with standard PyPI dependencies, it installs with a single pip install on top of any Python-based base image.

Minimal Dockerfile

FROM python:3.12-slim

WORKDIR /app
RUN pip install --no-cache-dir pyloseq

COPY analysis.py .
CMD ["python", "analysis.py"]

Common base images

All of the following work with a plain pip install pyloseq:

Image	When to use
python:3.12-slim	Smallest footprint for production pipelines
python:3.12	Full Debian base; useful when you also need compiled system tools
continuumio/miniconda3	When the rest of your environment is conda-managed
jupyter/scipy-notebook	Interactive notebooks with JupyterLab already installed
ubuntu:24.04	System-level dependencies managed by apt; install Python via apt-get install python3-pip first

Conda environments

If your base image uses conda, install pyloseq from PyPI into the base or a named environment:

# into the base conda environment
conda run pip install pyloseq

# or into a named environment
conda create -n microbiome python=3.12
conda run -n microbiome pip install pyloseq

pyloseq is not currently on conda-forge. All dependencies that have conda-forge packages will be resolved from there by conda when you later call conda install; pyloseq itself is fetched from PyPI.

Verifying your image

A smoke test to confirm the install is working:

docker run --rm <your-image> python -c "
import pyloseq
from pyloseq import Phyloseq, OtuTable
print('pyloseq', pyloseq.__version__, 'OK')
"

The scripts/test_container_install.sh script in the repository runs this test across all supported base images.