Use Conda
All pipeline software should be packaged using either Bioconda or conda-forge.
Using conda for software packaging is strongly preferred because:
- Many pipeline users may be unable to use Docker / Singularity and run using Conda
- Conda packages are automatically packaged into Docker / Singularity images using biocontainers (Bioconda) or Seqera Containers (bioconda and conda-forge).
- We have automation to detect new bioconda / conda-forge software releases, which can trigger automated nf-core module updates
Note that all Bioconductor R packages are available via Bioconda.
Seqera Containers makes it easy to bundle multiple conda packages together into a single container, as well as providing images for both aarch64 and arm64 CPU architectures.
Anyone can package software tools into Bioconda / conda-forge - you do not need to be the original author of the software. Many nf-core developers have packaged software for conda, which benefits the wider bioinformatics community.
To get help with packaging on conda, please see the #bioconda slack channel.