popup link content for phylotrees

[adf-ncgr]

decide which links to include/exclude from the popup menu on the phylotree

[LEGUME-179] created by adf_ncgr

[adf-ncgr]

For early link targets, I would vote for
1) Phytozome (since there is good coverage except for Arachis)
This probably requires hashing to the Phytozome ID, i.e. Medtr3g108450.1 is at gene page
http://phytozome.jgi.doe.gov/pz/portal.html#!gene?search=1&detail=1&method=4432&searchText=transcriptid:31056999
2) SoyBase for soybean (this would make the gene families much more useful there)
3) The respective LIS and PeanutBase browsers. One method (Nathan may have refinements for the API):
http://lis-stage.agron.iastate.edu/gb2/gbrowse/Mt4.0/?q=Medtr3g108450.1
http://peanutbase-stage.agron.iastate.edu/gb2/gbrowse/Aradu1.0/?q=Aradu.L9R8I

by scannon

[adf-ncgr]

Regarding the links to GBrowse: if it's not too onerous, it's greatly beneficial for performance and (conceivably) beneficial for specificity to restrict the searches to the genes track for each species. As we store each track in a separate database, a search can be restricted in this manner by adding the appropriate dbid to the search; e.g.:

http://lis-stage.agron.iastate.edu/gb2/gbrowse/Ca1.0/?q=Ca_28062;dbid=cicer_arietinum_gene_models

This dbid differs between each GBrowse instance; here is a list

legumeinfo:
Ca1.0: cicer_arietinum_gene_models
Cc1.0: gene_models
Lj3.0: gene_models
Mt4.0: gene_models
Pv1.0: gene_models

peanutbase:
Aradu1.0: gene_models
Araip1.0: gene_models

soybase:
gmax2.0: gene_models_wm82_a2_v1

by nathanweeks

[adf-ncgr]

OK, I have implemented some of the suggested linkouts for lis-stage (just modified per Nathan's suggestion).
Currently the way Alex's code is structured it seems easiest to have a single external link target for a given species
(probably it won't be too hard to change this later when there is time to think), so I have provisionally done:
glyma -> soybase gbrowse
phavu -> phytozome gene page
medtr -> JCVI jbrowse
aradu/araip -> peanutbase

and for the non-legumes, currently only arath -> AIP

NB: the arachis links are currently set to the -stage servers to facilitate your test-driving, but WILL HAVE TO BE CHANGED PRIOR TO ROLLOUT. It seems likely there's a good way of having non-relative links that are context-sensitive (dev/stage/production), but
that will have to be done later.

let me know if you'd like to change anything...

by adf_ncgr

[adf-ncgr]

Andrew: can we use relative URLs? Most (all?) should work for the production instances as-is, and those that don't can be hand-massaged for testing. I'm concerned that nobody is going to remember to change the links before lis-stage is "productionized".

by nathanweeks

[adf-ncgr]

I could for legumeinfo targets, but don't think I can do that for peanutbase targets, unless you can advise on an approach;
I definitely share your concern!

by adf_ncgr

[adf-ncgr]

Oh right---in that case, you can just use absolute URLs to peanutbase.org, and we can change the URLs by hand when testing (peanutbase.org/gb2/gbrowse/Aradu1.0 -> peanutbase-stage.agron.iastate.edu/gb2/gbrowse/Aradu1.0x).

by nathanweeks

[adf-ncgr]

I see your point that it's better to make people testing do the manual work than rely on our collective memory to change the code
before rollover; I'll change it now....

by adf_ncgr

[adf-ncgr]

FWIW, one approach that we could attempt in the future is to use a relative URL (e.g., "/peanutbase") that Apache redirects to peanutbase-stage.agron.iastate.edu or peanutbase.org depending on whether the hostname is lis-stage or legumeinfo.org. By reducing the dev/stage/production-specific differences to the Apache config file, we could potentially make alternate code paths for dev/stage/prod elsewhere unnecessary.

by nathanweeks

[adf-ncgr]

sounds worth exploring to me

by adf_ncgr

[adf-ncgr]

Looking good to me. I have tested genes from glyma, medtr, aradu, araip, and arath (which spell-check wants to change to glimmer, meddler, arid, aria, and wrath, respectively. Talk about meddling.).
For testing, I don't mind manually replacing e.g. peanutbase.org with
http://peanutbase-stage.agron.iastate.edu/

I'll provisionally close this issue (re-open if necessary).

by scannon