Ten simple rules and a template for creating workflows-as-applications

As bioinformatics analyses increase in size and complexity, workflow managers are becoming more popular for building pipelines. Workflow managers, such as Snakemake , Nextflow, and Cromwell with WDL or CWL, empower researchers to build robust pipelines that call a series of tools and scripts to perform a bespoke analysis. Workflow managers enable non-bioinformaticians to run published pipelines with confidence, and workflow managers with graphical user interfaces such as Galaxy and BioWorkflow have helped non-bioinformaticians create their own simple pipelines. Earlier tools for workflow management have been around for a while, including GNU Make, ruffus, doit, rake for ruby, and Makeflow. However, the integration of cluster and cloud computing support in Snakemake, Nextflow, and Cromwell helped drive their current popularity. The use of workflow managers facilitates following the FAIR (Findable, Accessible, Interoperable, Reusable) guiding principles for open scientific research. Interestingly, many existing bioinformatics command line tools are wrappers for a series of other software, but since that is the goal of workflow managers, they can be used instead. Examples of command line tools built on a workflow manager include Hecatomb, ATLAS, VirSorter2, spacegraphcats, BlobToolKit, and PGAP. These tools all consist of two key components: a convenience launcher, which provides the command line interface for the tool and compiles the configuration from user command line arguments, and the workflow pipeline and associated files, which performs the actual analysis...