New curation type "protein localization dependency"

[pombase-admin]

New curation type "protein localization dependency"

This is a slightly odd one. We would like to have an annotation type "protein localization dependency" This would be identical to the phenotype annotation, but would only show the terms in the phenotype ontology below "protein localization", and would have a slightly different text. Explaining that "localization dependencies" are captures as phenotypes of the mutated gene, with the protein whose localization is being monitored added in the "optional information section" (and later will be replaced by an annotation extension)

We think this would encourage the community to curate 'localization dependencies" which they don't seem to think of as "classical phenotypes"

(I thought this was already on the tracker but I can't see it)

This may require https://sourceforge.net/p/pombase/curation-tool/428/

Original comment by: ValWood

[pombase-admin]

This is another case where it would be good to get Oort/OWLtools to do the processing and report the sub-set of GO that we need.

One question: which relations should we follow to get the descendants of "protein localization" for this curation type? I can see three for the direct descendants: is_a, part_of and regulates.

Original comment by: kimrutherford

[pombase-admin]

I don't know...I'm referring to FYPO localization therms here, not GO (this is a phenotype subset)....I don't know what the realtions in FYPO are but does the Q. still stand?

val

Original comment by: ValWood

[pombase-admin]

I was confused because you say: phenotype ontology below "protein localization"

I looked up "protein localization" and the only term I found with that name was in GO, so I assumed you meant GO.

Which FYPO term did you mean?

There are lots of "relationship: ..." lines in the fypo-simple.obo file (here shown with the number of occurences):

  2 relationship: inheres_in
  3 relationship: results_in_formation_of
 11 relationship: part_of
 13 relationship: has_output
 37 relationship: includes_cells_with_phenotype
 47 relationship: comprises_cells_with_phenotype
 49 relationship: output_of
261 relationship: has_part

Which ones should we follow to get the sub-set you're interested in? We probably won't encounter all of those relationship as descendants of the FYPO term.

Original comment by: kimrutherford

[pombase-admin]

my mistake

it would be "abnormal protein localization I think" (or maybe it should be a 'grouping term' protein localization phenotype, which does not exist right now.... will discuss with Midori, but for starters "abnormal protein localization will work" I am not sure what relationships are under here.

Original comment by: ValWood

[pombase-admin]

Val did me a ticket on the fypo tracker for a top-level "protein localization phenotype" term, so I'll add it next time I'm editing the ontology anyway. Might be this week, or next.

Once that's done, I think you'll only need to follow is_a and output_of. Almost all the links between terms of interest are is_a; there's one output_of (FYPO:0000782 output_of FYPO:0000443) in this area at present, and I might add a few more eventually. Maybe also has_output in future? but none yet, and I'll try to remember to warn you if I do add any later.

Original comment by: mah11

[pombase-admin]

There's now a shiny new term to use:

protein localization phenotype FYPO:0002333

Original comment by: mah11

[kimrutherford]

Should be easy to implement after #1023.

[kimrutherford]

This should be straightforward now. How urgent is this new curation type?

[ValWood]

Not super urgent. I grouped with 'new data types' which I think comes below things like 'file export' and 'filtering related tasks' in the current priorities list (mainly because we don't curate a lot of these, and we can do them in artemis for now...)

[ValWood]

Low priority, I think some of these would be helpful eventually, but there might be better ways to do this. As time goes on people seem to be finding these terms more anyway, probably a they see them on pages ....

[ValWood]

This might be useful to do, even though people are now much better at making these annotations.