GalaxyAddTool

How-to for adding a new tool into UAB's Galaxy instance. This details adding a tool that has it's own executable(s), rather than a pure python script.

pre-requisits

1. Install the tools on the cluster
   • /share/apps/your_tool (email mhanby@uab.edu to get the directory created)
   • /lustre/project/your_project (email mhanby@uab.edu to get the directory created)
   • /share/apps/galaxy/galaxy-tools (only if this will ONLY be used in Galaxy)
2. Create a personal galaxy for developement & testing
   • See Shantanu's Create a personal galaxy page.
3. Add the executable path to the galaxy environment
   • $galaxyroot/modulefiles/galaxy*: edit the module definition file to add your tool location to the path
4. Load the updated module and make sure the command works from the commandline!
   • module load galaxy
   • test, test, test...

Special Case: REST or SPARQL endpoints

There are already tools for integrating these through a wizard. See the Galaxy Toolshed and navigate to Tools Shed -> Data Source

Creating a command-wrapper by hand

Galaxy's public instructions for adding tools: http://wiki.g2.bx.psu.edu/Admin/Tools

Note: If you only need a simple wrapper for an exiting program, you'll only need to create an XML wrapper to describe it (my_tool.xml). However, if you need to create temporary directories or do any other pre- or post- processing, you'll also want to create some adapter code, and have my_tool.xml call that, rather than calling your tool directly. For example, sam_pileup.xml calls sam_pileup.py, which sets up a temp directory, indexes genomes on the fly, etc., then calls "samtools pileup ..."

1. (optional) Create a my_tool.py/pl/etc adapter to do pre-/post- processing
2. Create a my_tool.xml wrapper to
   • define tool ID and display name
   • declare inputs
   • declare outputs
   • list test files
   • user documentation
3. Edit tool_conf.xml to add a line for the new XML file
4. Restart galaxy
   • ./run.sh --stop-daemon; ./run.sh --daemon