Pipeline Requirements
All released WARP pipelines should meet the following criteria:
- Pipelines are written in WDL, modular, and optimized for Google cloud. The pipelines are often portable to other environments.
- Pipelines have plumbing (fast) tests and scientific tests for catching any unintended changes in data processing.
- Pipelines are semantically versioned, have a changelog, and are packaged into a release with all dependencies.
- Pipelines are released to Dockstore automatically upon release and are available in the cloud-based Terra platform.
- Pipelines have example inputs alongside the pipeline.
- Pipelines have a Readme.md describing the pipeline.
- Pipelines use public docker containers and only open source tools.
- Pipelines are developed in a collaboration between software engineers and a scientific owner of each pipeline who must approve all changes to the code and verify that any resulting changes in outputs are scientifically valid.
- Pipelines have a publication-style methods section to enable easy citation.
tip
Read more about our pipeline development in our Best Practices documentation.
When citing WARP, please use the following:
Kylee Degatano, Aseel Awdeh, Robert Sidney Cox III, Wes Dingman, George Grant, Farzaneh Khajouei, Elizabeth Kiernan, Kishori Konwar, Kaylee L Mathews, Kevin Palis, Nikelle Petrillo, Geraldine Van der Auwera, Chengchen (Rex) Wang, Jessica Way. "Warp Analysis Research Pipelines: Cloud-optimized workflows for biological data processing and reproducible analysis." Bioinformatics, 2025; https://doi.org/10.1093/bioinformatics/btaf494