Genomics in the Cloud

Chapter 9. Deciphering Real Genomics Workflows

In Chapter 8, we showed you how to string commands into workflows by using the WDL language, and we had you practice running those workflows on your cloud VM using Cromwell. Throughout that chapter, we used fairly simple example workflows in order to focus on the basic syntax and rules of WDL. In this chapter, we switch gears and tackle workflows that are more complex than what you’ve seen so far. Rest assured, we neither expect you to instantly master all of the intricacies involved nor to memorize the various code features that you’ll encounter.

Our main goal is to expose you to the logic, patterns, and strategies used in real genomics workflows and, in the process, present you with a methodology for deciphering new workflows of arbitrary complexity. To that end, we’ve selected two workflows from the gatk-workflows collection, but we won’t tell you up front what they do. Instead, for each mystery workflow, we tell you the functionality you’re going to learn about and then walk you through a series of steps to learn what the workflow does and how the functionality that we’re interested in is implemented. This won’t turn you into a WDL developer overnight, but it will equip you with the skills to decipher other people’s workflows and, with a bit of practice, learn to modify them to serve your own purposes if needed.

Mystery Workflow #1: Flexibility Through Conditionals

Our first mystery workflow gives you the opportunity to learn ...