More about running fastq-to-gVCF¶

When we do releases, we run a test pipeline that is intended to be representative of a bioinformatics workload.

1. Deploy the version of arvados-cwl-runner that you want to test and make sure that the corresponding arvados/jobs image has been built and uploaded to docker hub or built using the arvados/build/build-dev-docker-jobs-image.sh script and uploaded using arv-keepdocker.

2. Clone https://git.arvados.org/arvados-tutorial.git/

3. Create an Arvados project for the test run

4. cd arvados/tutorial/WGS-processing

5. Run the following command: arvados-cwl-runner --no-wait --disable-reuse --project-uuid <my project> cwl/wgs-processing-wf.cwl yml/wgs-processing-wf-chr19.yml

6. Monitor this for success. It usually takes about an hour to run.

If you are running this on pirca then all the data should already be present. If you are running it from somewhere else, you may need to do some additional data copying from pirca to the other cluster. The input document yml/wgs-processing-wf-chr19.yml has the portable data hashes of the collections.