search

PHI-base / curation

PHI-base curation
0 stars 0 forks source link

Restricted 'do not annotate' terms used in some sessions #85

Closed jseager7 closed 3 years ago

jseager7 commented 3 years ago

(References https://github.com/pombase/canto/issues/2255)

Not sure if this has already been discussed, but according to the syntax errors in our GAF file, two curation sessions have annotations that use a GO term tagged with 'gocheck_do_not_annotate'. They are:

ValWood commented 3 years ago

'transcription regulator activity If we don't know the mechanism of transcription regulation, (i.e corepressor, GTF, DNA-binding TF), we should change this annotation to "+/- regulation of transcription"

intracellular anatomical structure Presumably it should be possible to make this one more precise.

jseager7 commented 3 years ago

we should change this annotation to "+/- regulation of transcription"

There doesn't seem to be an exact match for 'positive regulation of transcription' (or negative) in GO. Instead there are qualified versions like 'viral transcription', 'reverse transcription', 'mitochondrial transcription', etc. Here's the OLS search results for 'positive regulation of transcription':

https://www.ebi.ac.uk/ols/search?q=positive+regulation+of+transcription&groupField=iri&start=0&ontology=go

Did you have a particular term in mind?

mah11 commented 3 years ago

There doesn't seem to be an exact match for 'positive regulation of transcription' (or negative) in GO.

The relevant terms have slightly more complicated names: GO:0045892 ! negative regulation of transcription, DNA-templated GO:0045893 ! positive regulation of transcription, DNA-templated

The "DNA-templated" is one of the ways GO distinguishes "forward" transcription from reverse transcription.

ValWood commented 3 years ago

This might also be able to be more specific (polymerase II?) @CuzickA

CuzickA commented 3 years ago

@ValWood can we have a quick look at these when we chat on Monday

Why is PHI-Canto letting us annotate with GO 'do not annotate' terms?

ValWood commented 3 years ago

I'm not sure if we have a hard check for this (I thought we did, but it seems not)

@mah11 do you remember? I thought at one point if a "do not annotate" restriction was present you had to select a child term?

I tested "cell cycle" which has this comment: Restrictions This term should not be used for direct manual annotation; it may, however, be used for mapping to external vocabularies in order to create electronic annotations.

(maybe thre are different tyoe of commet?)

kimrutherford commented 3 years ago

@jseager7 here is the PomBase config: https://github.com/pombase/pombase-config/blob/master/canto/pombase/production_canto_deploy.yaml#L95

jseager7 commented 3 years ago

Turns out this was due to 'gocheck_do_not_annotate' and 'gocheck_do_not_manually_annotate' being missing from the PHI-Canto configuration file, so the GO terms weren't being ignored when they should've been. I'll fix the config file then close this issue when I've confirmed that the fix has worked.

jseager7 commented 3 years ago

The change is active on the demo server now and should be active on the main server tomorrow, but I'll leave this issue open until the GO terms above have been replaced with valid terms.

ValWood commented 3 years ago

~this is confusing, when I enter the PMID I get: Welcome to Canto at PomBase You are about to start curating data from this paper. You can either curate this paper yourself, or send it to a member of your lab to curate. which implies that it isn't curated?~

I'll open a ticket about this.