search

pombase / canto

The PomBase community curation tool
https://curation.pombase.org
Other
19 stars 7 forks source link

config file for protein-protein interaction phenotypes #2044

Closed ValWood closed 5 years ago

ValWood commented 5 years ago

This needs an extension configuration. All branches of PHIPO (pathogen, host and interaction) phenotypes which mention pathogen host interactions should prompt for 2 assayed using genes.

I'm not sure if this needs to be configured in the Canto settings, or the cardinality in the assayed_using file is sufficient.

@mah11 can advise.

This is fairly high priority as we will use it a lot and it's a bit of PINTA to navigate back to the gene list to copy paste the uniprot IDs

jseager7 commented 5 years ago

Unfortunately I still haven't managed to get the assayed_using relation working on the PHI-Canto servers, even though I tried to follow PomBase's configuration setup. I'm not really sure what could be going wrong, so I'll have to discuss this with @kimrutherford, seeing as this is blocking multiple issues now.

kimrutherford commented 5 years ago

Hi James. Could you attach your configuration to the issue or add a link to it?

jseager7 commented 5 years ago

Here's the links to the config files (private repository, but you should all have access):

https://github.com/PHI-base/config/blob/master/annotation_extension/phipo_extensions.tsv

https://github.com/PHI-base/config/blob/master/annotation_extension/phipo_extension_relations.obo

https://github.com/PHI-base/config/blob/master/canto_deploy.yaml

kimrutherford commented 5 years ago

Thanks for the files. It all looks OK. It's worth checking that there are no errors that make the load script bail out. Can you also check the process exit with: echo $? ? It should print 0

On Monday when I'm back home I'll try loading them locally.

mah11 commented 5 years ago

There may be nothing for @kimrutherford to do -- now that I have read access to the extension config file, I can see that it has only specified an assayed_using configuration for one branch of the ontology (PHIPO_0001060 altered RNA level). I ferreted out an active session and tested, and the assayed_using extension option does work for terms in that branch.

So all you need to do is update the config. Try the one I'm uploading here (I don't have permissions to create a pull request; also ignore the file extension, as I had to add .txt to make GitHub upload it phipo_extensions.tsv-2019-10-03.txt).

ValWood commented 5 years ago

Thanks!

jseager7 commented 5 years ago

@mah11 Thanks for this. I'll test the config on our servers shortly.

With regards to the permissions problems, it looks like with Read permissions, you can only open a pull request on the config repository by forking the repository first. I've just given you Write permissions for now, for the sake of convenience. That should allow you to make a pull request on the repository itself.

jseager7 commented 5 years ago

It seems like the config started working once I restarted Canto on the server. Not sure if that's supposed to be required or not.

All of the extensions in the config file linked above (phipo_extensions.tsv-2019-10-03.txt) are now enabled on the demo server, and they should be enabled on the main server sometime tomorrow.

One thing I'm not sure about: @ValWood you said in the first comment that "All branches of PHIPO [...] which mention pathogen host interactions should prompt for 2 assayed using genes." Did you intend for the extension to apply to the entire 'pathogen-host interaction phenotype' branch? Or is the current term domain of 'protein protein interaction phenotype' (PHIPO:0000132) the correct domain for this extension? The issue title suggests that it was for the latter term.

mah11 commented 5 years ago

Did you intend for the extension to apply to the entire 'pathogen-host interaction phenotype' branch? Or is the current term domain of 'protein protein interaction phenotype' (PHIPO:0000132) the correct domain for this extension? The issue title suggests that it was for the latter term.

I hope Val meant the latter, because that would be correct, whereas any broader domain would be incorrect.

kimrutherford commented 5 years ago

It seems like the config started working once I restarted Canto on the server. Not sure if that's supposed to be required or not.

A restart is required if the config is changed.

jseager7 commented 5 years ago

This is enabled on both servers now, so I'm going to close this. Feel free to re-open if you notice any terms missing extensions.

ValWood commented 5 years ago

yes sory I meant 'protein protein interaction phenotype'