Hi there, I'm having several issues with the featured workspace involving a reproducible GWAS.
The WDL script fails for me, whether you use data you generated yourself from the notebook or the provided completed data. The null model part — whether you run it as provided or import the null model part directly from Dockstore — completes successfully. It is the second part that fails, and again, you get the same result whether you use it as provided or if you import the second part from Dockstore. In either case the error appears to be:
Error in `$<-.data.frame`(`*tmp*`, chr, value = character(0)) :
replacement has 0 rows, data has 2504
Calls: $<- -> $<-.data.frame
ln: failed to access '/cromwell_root/*.gz': No such file or directory
While debugging the above issue, I tried duplicating the GWAS notebook to regenerate the phenotypes file that the WDL workflow uses. This duplicated notebook hangs when it tries to import VCFs. I left it running over lunch and it seems it never progressed beyond that stage and the notebook just turned itself off, presumably due to inactivity. There was no error message.
In fact the lack of error messages applies elsewhere, as when debugging the hanging notebook I tried the command “os.getcwd()” and that also appears to have caused the notebook to not progress without an error message. The error itself isn’t the issue—I’m pretty sure it’s just because the os package wasn’t imported—it’s the lack of an error message. I don’t know if this is standard behavior for Jupyter, but it does seem like this will lead users to believe something is just taking a long time to run when it’s actually never going to complete.
Sorry if this is a bit of a tangle of bug reports and feature requests, but since they're all connected I figured I'd just make one big post. I have added the support group to the workspace as a writer as suggested in the stickied post.
Thanks in advance for your help!
Please sign in to leave a comment.