Hello all,
I'm looking at how best to incorporate Galaxy workflows into a publication - a standalone *.ga file in supplementary materials would probably work but is static and I fear likely to bit rot.
I think posting each workflow (or set of workflows) as a repository on the main Galaxy Tool Shed would be better, as I can continue to update it as needed, and it can be linking to it via the citable URL, e.g. Greg's example on the Test Tool Shed which declares some dependencies: http://testtoolshed.g2.bx.psu.edu/view/greg/heteroplasmy_workflow
(This assumes the Tool Shed URLs will be stable, but I'm having to assume that anyway for tool wrappers.)
However, my natural next question is how do we define unit tests for these workflows?
In the case of a methods paper, some sample data for testing would be reasonable. In the case of a more applied paper, the workflow could perhaps include the actual data used in the paper and reproduce the published results (ambitious but quite possible for moderate sized datasets).
I'm guessing this would be a major enhancement - something for Greg and Dave to tackle once the current round of work for testing tools on the Tool Shed is done (grin).
Shall I open a Trello issue for this?
Regards,
Peter