By Moran Cabili, product manager, Data Sciences Platform at the Broad Institute

We heard from many of you --both new FireCloud users and experienced WDL pipeline developers-- that you need to be able to quickly test that a WDL workflow can be run successfully on FireCloud. Until now, you were required to reference an entity in the workspace data model, which took extra effort and tended to confuse newcomers. We are happy to announce that this speed bump has been eliminated; you can now bypass the data model and even upload a JSON of inputs to get your WDL up and running in record time.

You don't need a populated data model to run a workflow

You can run your workflows with just your method and your inputs specified directly in the method config. Before you launch, just uncheck the checkbox ‘Configure inputs/outputs using the Workspace Data Model’ in the Method Configurations tab. If you’d like to continue using the workspace data model, leave it checked, and it will operate as usual.

Note that referencing the data model is still needed for launching submissions that include multiple workflows, for example when you want to run the same workflow on multiple samples in a sample set in parallel.

You can upload a JSON file of inputs to populate the method configuration

JSON is the format most commonly used to specify inputs for a WDL workflow outside of FireCloud, so this is especially useful if your workflow was developed and tested outside of FireCloud and therefore already comes with a JSON file of inputs. For example, this feature makes it easy for you to run a WDL from Dockstore if it is accompanied by a JSON file, as is usually the case (see the Test file section in the Dockstore repository of interest). Note that if you are trying to use a WDL with a JSON file that was tailored for a different platform, you'll need to make sure you update the paths to any files to point to locations in Google Cloud storage that FireCloud can access.

You can download a JSON file to copy and re-use lists of input files and parameters

You can now download your JSON file too, which comes in handy when you want to use the same inputs across method configurations. For example, if you create a new method configuration, normally you’d have to enter all of the inputs manually every time, even if this new configuration will use many of the same inputs as an existing configuration. Now you can download the inputs from the prior configuration as a JSON and upload that to populate the new configuration.

Admittedly it would be even better to be able to copy inputs from one configuration to the other within FireCloud itself --and we're looking at ways to make that happen in the future-- but at least the ability to save inputs to file and re-upload is a big step up from having to fill out each field one by one manually! Right?

So in short, we hope these new capabilities will make it easier for you to get your workflows up and running quickly, whether it's your first time using FireCloud, or you're an old hand testing new WDLs. We do however still recommend making use of the power of the data model for batch processing, especially when you have a lot of data to crank through your pipelines. In fact, we also added some convenience for when you choose to use the data model: you can now set the outputs of your workflow to get hooked up to attributes named after the corresponding output names used in the WDL.

To get you fully up to speed, we’ve updated our Quick Start Guide on method configurations, as well as the tutorials on launching an analysis and configuring a method in a workspace.

Return to top

Mon 6 Aug 2018
Comment on this article

- Recent posts

- Follow us on Twitter



If you missed it, Junko will be back at the Broad booth 3048 Monday at 9am — swing by for a FireCloud demo or to he…
1 Apr 19
Want to see FireCloud in action, and hear more about our upcoming upgrade to being powered by Terra? Come to our de…
28 Mar 19
RT @NCIDataSci: Looking to analyze #data without downloading large data sets? Use the #NCICloud Resources! These three resources support yo…
28 Mar 19
@jamesaeddy Going to phone a friend on that one — @geoffjentry care to explain?
19 Mar 19

- Our favorite tweets from others

Pipelines API is used by a number of popular tools, such as Firecloud/Terra from @broadinstitute. @BroadFireCloud
11 Apr 19
The macaque genome isn't finished so it has over 200K contigs. I call them the Rhesus pieces.
15 Mar 19
It's been years in the making, but 500k structural variants (SVs) from 15k genomes in #gnomAD are out 🎉today🎉! Pre…
15 Mar 19
@fdmts @tanyacash21 and R. Munshi @broadinstitute are beginning their session on hype vs. reality in cloud capabili…
12 Mar 19
watching "Genomics TV" #bioinformatics #GATK @broadinstitute
5 Mar 19

See more of our favorite tweets...