Be creative - create/import workflows

[bgruening]

We have a variety of tools and visualisations - we should create example workflows for them and cover use-cases.

Can everyone come up with at least on workflow?

[bgruening]

You can use the already existent Docker image to create a workflow and export it as *.ga file and submit it into this repo. it should be easy as this.

@s-will has already mentioned to have a few ideas as well.

[dyusuf]

@bgruening @s-will I am thinking to use the RNA-seq workflow which was used in our courses.

[bgruening]

Sounds good. Hopefully all tools are integrated. Can you also create one with one of your tools?

[bgruening]

@dyusuf: @jfallmann is working on tours for the main training material https://github.com/bgruening/training-material/blob/master/RNA-Seq/tutorials/ref_based.md I guess it would be nice if we can have a workflow for this one. So that everything is coherent.

[jfallmann]

@bgruening @dyusuf Think that's a good idea, the tour should be easy to adopt for other datasets later if that's needed

[dyusuf]

@bgruening @jfallmann sure, I do so

[dyusuf]

@bgruening

I try to use subworkflow. Well, after adding a subworkflow, there are no options available for it even though it works fine as an independent workflow.

any hints?

[dyusuf]

@bgruening the subworkflow only works when "copy and insert individual steps"

[bgruening]

@dyusuf let's use the copy and insert for now. I think this is an upstream bug. @s-will are you working on a small example ViennaRNA workflow?

[dyusuf]

@bgruening for mapping, how to resolve the issue for the large data like genome.

usually, a galaxy server has built-in index...

[bgruening]

We could include the reference genomes as well ... but this would blow up the image. Can we give a link to the fasta reference file (UCSC link) to download as well?

[dyusuf]

@bgruening

when I using tophat to do mapping the outputs are empty without errors.

using the same datasets, the Freiburg server generates the outputs with sizes.

could you do a simple mapping with any datasets to see whether the issues can be reproduced.

ps: the reference is not built-in but from history

[dyusuf]

@bgruening

please look at the following two examples. It seems tophat may have the issue on workbench.

workbench, empty outputs: http://192.52.2.98/u/admin/h/unnamed-history

freiburg, outputs with sizes: http://galaxy.uni-freiburg.de/u/dilmurat-yusuf/h/testtophat

[bgruening]

I guess this is because tophat needs more memory and we need to change the job_conf.xml

[dyusuf]

@bgruening let me know when it is fixed so I can test the workflow.

[dyusuf]

@bgruening I want to integrate dexseq into the workflow, while the installation of the galaxy tool invokes the old dependency resolver where are some problems with the dependencies. I am wondering if there is a version that relies on the conda pkg.

[bgruening]

@dyusuf I think we have a conda package for it, but I never found the time to update the wrapper :(

[dyusuf]

@bgruening I am trying to update the dexseq wrappers to use the available conda pkgs.

so far, it went well with dexseq.xml.

for dexseq_count.xml, an environmental variable $DEXSEQ_ROOT is required to access two python scripts in the Dexseq package.

https://github.com/galaxyproject/tools-iuc/blob/643293a896a2ccfac10fe995f48c7f01c1a89a7f/tools/dexseq/dexseq_count.xml#L19

with the old resolver, the variable can be set with set_environment which seems being ignored by the conda resolver.

Do you know how to resolve this issue?

[bgruening]

I guess the correct solution is to put the python scripts into PREFIX/bin in the conda package?

[dyusuf]

@bgruening yes for sure, that can be done at the level of conda package. I was just wondering any solution from Galaxy so that I need not go to bioconda to fix the recipe :) well, it seems the only solution.

by the way, I found the bioconda recipe is for version 1.18.4 while the conda package is 1.16.6. do you know why the 1.18.4 version was not compiled.

https://github.com/bioconda/bioconda-recipes/blob/master/recipes/bioconductor-dexseq/meta.yaml

https://anaconda.org/bioconda/bioconductor-dexseq/files

[bgruening]

Probably, just a lack of time on my side :( There is a hacky way on the Galaxy side, involving which or find ;)

[bgruening]

Thanks Dili!

[dyusuf]

@bgruening regarding the tophat failure, the following error might be the culprit,

/export/galaxy-central/database/job_working_directory/000/36/conda-env/bin/tophat2: /export/tool_deps/_conda/pkgs/tophat-2.1.0-py35_0/bin/tophat: /opt/anaconda1anaconda2anaconda3/bin/python: bad interpreter: No such file or directory

this might be the issue related to the conda pkg.

[dyusuf]

@bgruening I did planemo test on the galaxy tophat2 tool (in my ubuntu), there was no error with conda dependencies. then is it possible something wrong with docker setting in relation to conda?

[bgruening]

According to @dpryan79 tophat never worked on python3, so the shebang is probably python2 in this script, which is not available? https://github.com/bioconda/bioconda-recipes/pull/3513

[dyusuf]

@bgruening so what would be the plan to make tophat work? a workbench without this aligner does not sound promising.

[bgruening]

We should fix it ;) Can you try again, hopefully it is fixed :)

[bgruening]

@s-will can you get one workflow/tour up and running?

[bgruening]

@yhoogstrate any small workflow possible. With a nice description maybe to get a VIs after your WF? @TorHou can you also please add one. @bagnacan can the Rostock group please also add one or two ... preferably with the tools already included.

[yhoogstrate]

Are we talking about Rna 2d structure and which deadline? I won't be able to find Time this weekend