descriptions of different GO terms are merged in ggs_graph

[thomas-keller]

Thanks for developing this great package.

I'm working on converting some of these functions for use with the msig database, and noticed that in the ggs_graph call, the GO expanded descriptions are merged. You can see this on the user guide (GeneTonic User guide). All the terms have the same (large) description because they've been collapsed together.

line 200 in
ggs_graph
should go from
paste0(strwrap(enriched_gsdescs[nodeIDs_gs], 50), collapse = "<br>")

to something like
sapply(enriched_gsdescs[nodeIDs_gs],function(x) paste0(strwrap(x,50),collapse='<br>'))

If there's interest in the msig stuff, I can work on a pull request later. With the shaker function I made for GSVA, it seems to mostly work for the few functions I've focused on so far (ggs_graph,enrichment_map). Some functions would require some changes, like the assumption of a DEseq results input (I'm using limma right now with gsva).

Let me know if you'd be interested in adding msig functionality, I can do a pull request eventually.

[federicomarini]

Hey Thomas,
thanks for the kind words! Very glad you enjoy using it.

You are referring to the content of the tooltips, if I got your point right? (the link in the email pointed to L325, which is very odd - and something else...).

Aaaanyway: great catch, thank you for spotting this AND for providing a fix to that.
This also shows up btw also with shorter GO terms (I initially thought it was only for longer descriptions...).
Basically it was a naif (read, "incorrect but innocent") usage of the almighty paste0 function.

I will commit & push the fix!

As for the other aspect: Sure, if you want to put together a PR, I am happy to review it!
Probably you can have a look at the DEformats package to interconvert among DE results formats?

Federico