galaxy celikalper84  2017-11-02 08:33
Status: Closed
 
Hi,

I am trying to get LabKey and Galaxy to communicate so that we can leverage the NGS pipelines that are already built in Galaxy. Currently I have local instances of both running in my computer (for test and development purposes). From the documentation I believe there used to be some sort of communication between the two platforms but cannot find any more details about it.

Thanks for all the help.
Alper

Here are some details:
LabKey version 17.2
tomcat version 8.5.23
Postgres version psql 9.6.5
Galaxy Version 17.09
OS Ubuntu 16.04
 
 
adam responded:  2017-11-02 12:13
Hi Alper,

As you've probably seen, the genotyping module includes the ability to generate files containing reads and reference sequences of interest from data imported and curated in LabKey, pass those files to Galaxy, invoke a Galaxy workflow that performs alignment, and load the matches back into LabKey. This workflow was developed 5+ years ago and tailored for a specific lab's needs; it hasn't been generalized or updated to the latest version of the Galaxy HTTP API. But, this might be a reasonable starting point if you wanted to develop integration between LabKey and your NGS pipelines. (Or LabKey could work with you to develop this integration.)

The other possibility is the "galaxyintegration" module developed by Ben Bimber at OHSU. In the SVN repository: /externalModules/DISCVR/galaxyintegration. I'm not very familiar with the feature set this module provides.

Adam
 
Jon (LabKey DevOps) responded:  2017-11-02 12:21
There is also another thing Ben Bimber made as well as an alternative to Galaxy:

https://github.com/bbimber/discvr-seq/wiki
 
Ben Bimber responded:  2017-11-02 14:46
Hi Alper,

To give a little more detail on those: we use DISCVR-Seq and associated modules to manage raw sequence data and kick off various alignment and analysis pipelines across many servers/projects, and there are a few groups I know about outside our institution also using it. It's designed to primarily manage your data and outputs, which we never found Galaxy to do very well. It has some flexibility in defining pipelines, but it's not designed to be anything like Galaxy on that front. We're actively continuing to develop and expand this.

I also wrote a module (galaxyintegration), which is also part of LabKey's SVN. This was really just a proof of concept module for a multi-site collaborative project. It uses an existing java library (blend4j) to port the outputs from a galaxy run into LabKey's experiment/pipeline system. Crudely speaking, it exports the locations of output files and record of steps performed from galaxy and stores those within LabKey, so they can be viewed as though it was a pipeline run via LabKey. Given we dont use Galaxy, I dont think this has ever been used seriously, but it should work. It might provide a model for how one might make a tighter integration between the two (basically leveraging blend4j and Galaxy's HTTP APIs); however, it's nothing I expect us to build out.

-Ben