The BSMN common workflow implemented in the Common Workflow Language.
This pipeline is an implementation of an existing pipeline available at https://github.com/bsmn/bsmn-pipeline.
This implementation requires a CWL interpreter to run. The most basic of these is the CWL reference implementation, available from https://github.com/common-workflow-language/cwltool. See the documentation on that page for installation, which requires Python.
This implementation currently requires data files to be stored and index in Synapse. It uses fastq files as inputs, which need to be index in a table or file view and annotated with specimenID, readPair = [1,2], and fileFormat = 'fastq'.
It also requires a reference genome (prefix file for BWA) to be accessible in Synapse. This requires a Synapse account.
- Download or clone this repository, and change to the directory that it is stored in.
- Modify the file test-workflow-entrypoint.json file to point to your own Synapse configuration file.
You can now run:
cwl-runner workflow-entrypoint.cwl test-workflow-entrypoint.json
The main workflow starts in workflow-entrypoint.cwl which gets initial files and queries Synapse for which samples and data files to process. The samples and associated files are then 'scattered' to run in parallel through a sub-workflow called synapse-bsmn-common-workflow.cwl. You can see the current steps implemented in the image below. You can generate this image using graphviz as such:
cwl-runner --print-dot workflow-entrypoint.cwl test-workflow-entrypoint.json | dot -Tpng > bsmn-common-workflow.png
Tests can be run with the cwltest tool, using this command:
cwltest --test test-descriptions.yml --tool cwl-runner