Is there something comparable to scatterCount in Queue?

I'd like to run gatk tools with data parallelism. Looking at tutorial, I am not sure if it comes with data parallelism for one sample.
I wonder if it is possible to split data into multiple pieces in wdl/cromwell by setting a number as Queue script splitting data using the scatterCount.
Would I need to implement a function getting genomic intervals from the count to do that? Thanks!

Best Answer

Answers

  • dayzcooldayzcool Member

    Thank you for your kind advice. Ah.. it would have been convenient if preset intervals don't need to be provided.

    I am comfortable and happy with Queue but I just begin to consider using wdl/cromwell as wdl looks to be preferred.
    May I ask what the killer feature of wdl/cromwell aside from learning curve? I just wonder what made you develop it.

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin
    @dayzcool Mainly we needed something that would give us more flexibility to run on different platforms, including multiple cloud environments.

    Some of Cromwell's features are still being actively developed to take advantage of the design features of WDL. We're also planning to expand the standard library of WDL functions to take care of things like indices automagically -- but it's still a while down the road because we have some production priorities to address first.
  • dayzcooldayzcool Member

    Thanks again for your comment. It was helpful to know the roadmap of the wdl.

Sign In or Register to comment.