This repository includes workflows for upstream processing of ATAC-seq data. The workflows are written in CWL and they are ready to run with a containerization solution like docker or sigularity.

Following workflows are available in the directory ./CWL/workflows:

• ATACseq.cwl - works with paired end ATAC data
• ATACseq_aligned.cwl - works with paired end ATAC data starting with already aligned reads

They depend on subworkflows located in ./CWL/workflow_modules and CommandLineTool wrapper located in ./CWL/tools.

Clone this repository, or download and extract a specific release.

You need to install a cwlrunner. We recommend cwltool for execution on a single machine or toil if you would like to use an HPC cluster or cloud infrastructure. A snapshot of a dedicated conda environment can be found at conda_toil_env.yml. However, every other runner that is fully compatible with CWL v1.0 should work, too.

To use the docker containers specified in the workflows, you need to set up docker or singularity.

Moreover, you need to install node.js for evaluation of the built-in java script expressions.

For workflow execution with cwltool and toil, you can find example shell scripts in the top directory (run_*.sh). They can be use in the following fashion (however, you likely need to slightly adjust the cwlrunner command line specified in the scripts):
bash ./run_cwltool.sh <cwl_workflow> <input_file_or_dir> <output_dir>
Or: bash ./run_toil.sh <cwl_workflow> <input_file_or_dir> <output_dir>

• <cwl_workflow> is a path to a cwl workflow.
• <input_file_or_dir> is the path to a YAML file containing input parameters, or the path to a directory with multiple YAML jobs that should be executed
• <output_dir> is the directory where final output files will be collected

Monitoring:
A log file (the name is ending with .yaml.log) will be created for each sample/run. You can open them to check the progress.

Example/test run:
Please try to run the following tiny test example to check whether your system is correctly configured. bash ./run_cwltool.sh ./CWL/workflows/ATACseq_pipeline.cwl ./tiny_test/test_main.atac.yml ~/cwl_test

Terminating a job:
If you like to terminate a specific run and all its subprocessed, you can find the corresponding job-ids in the file cwl_background_job_ids.log. Run: kill <job_id>.

Once the processing finished, you can find one QC report per workflow iteration (name ends with *_report.html). These reports were created using multiqc.

They can give you a deep understanding of your data's quality. However, you should also keep in mind that unexpected results can come from bugs in the custom scripts or the tools which are applied by the workflow. Therefore, check the report carefully. Especially, examine the "read flow" through the pipeline using the fastqc output: pay attention whether you encounter an unexpected loss of reads.

There are tiny synthetic tests located in ./tiny_test/. They are perfect for initial testing and debugging.