Using Ensembl gene annotation with Galaxy

Question: Using Ensembl gene annotation with Galaxy

15 months ago by

I'm trying to use ensembl gene annotations for my tool* but I don't understand how to map them to the UCSC dbkeys. There are multiple ensembl releases for each UCSC dbkey. My loc file has the latest ensembl release for each genome assembly/patch and I've tried to map it to the corresponding UCSC dbkey with this loc file:

When I try to run my tool with it, I see the full list from the loc file, but I get an error when executing: An invalid option was selected for builtin, 'GRCh37.p13', please verify. I don't understand where it even gets that value from.

*because building an ensembl annotation by release # is reproducible, unlike UCSC which always returns the latest gene annotations

ensembl ucsc galaxy dbkey • 541 views

ADD COMMENT • link •

written 15 months ago by matt.chambers42 • 50

15 months ago by

matt.chambers42 • 50

matt.chambers42 • 50 wrote:

It seems that I only got this error when overriding the preselected value from the workflow. When I edited the workflow to select a different value, the workflow submitted properly. This is with Galaxy 17.05.

ADD COMMENT • link written 15 months ago by matt.chambers42 • 50

This was your own tool? Do you consider this resolved?

Or, have you noticed this behavior with other tools (runtime database changes triggering errors)?

I'll test a few tools as well to see if I can reproduce. If there is a workflow functionality problem we'll want to fix it.

ADD REPLY • link written 15 months ago by Jennifer Hillman Jackson ♦ 25k

Yes, now I know where the bogus value came from. It was in my workflow .ga file that I imported. That value no longer existed in the loc file. But even when I overrode the value in the workflow run options, it still said invalid option. So it does seem like a UI bug. You should be able to reproduce by editing a .ga file and putting in a bogus value for a genome build parameter (then importing that workflow, running it, and trying to override the bogus value with a valid one).

ADD REPLY • link modified 15 months ago • written 15 months ago by matt.chambers42 • 50

Ok, the explanation helps. I wasn't able to reproduce a problem when switching databases during workflow execution.

I am not sure how often this would come up but tracking the problem seems worth it, as it is confusing to have a genome that is not the selected genome (at runtime) tossing the error. Would you like to open an issue ticket for this use case? This is the correct repository: https://github.com/galaxyproject/galaxy/issues.

Thanks! Jen, Galaxy team

ADD REPLY • link written 15 months ago by Jennifer Hillman Jackson ♦ 25k

Similar posts • Search »