Partial gather syntax in WDL?
I'm looking to implement a workflow wherein input samples start from fastq, get aligned, have variants called, and are then compared on the basis of variants calls.
In my minds eye, I figured using scatter syntax would make sense, since I could start with fastqs corresponding to my samples and then pull of alignment and variant calling in parallel.
My issue is that when I consider the final variant comparison stage, I have no clear way of setting up successor tasks that can takeover when their respective variant call steps are done.
For instance, say samples A-Z are setup to go through to variant calling; I want a setup that would let the comparison A-vs-C start when A and C are done, without waiting for the other tasks in the scatter to complete. If I was working directly on my grid with simple bash scripts, I'd setup A-v-C to succeed the A and C variant callings, but I think I'm misunderstanding how to handle the same scenario under WDL.
Can anyone please guide me on how best to set this up?
Comments
1 comment
Hi Pranay Vishwanath,
Thanks for writing in. Unfortunately, this support channel is only equipped to answer questions related to the Terra platform. We can take a look and help you interpret why your workflow failed in Terra, but determining how to write or optimize your WDL script is better achieved by doing it on your own (see Terra Support Policy). I would suggest reviewing the WDL GitHub spec, or posting your question to the expert community at bioinformatics.stackexchange.com.
Best,
Samantha
Please sign in to leave a comment.