AnnotationExtensions

== Implementing "Annotation extensions" in the curation tool

1. GO terms which can support an evidence extension will be documented in a config file:

Proposed config file format (tab-delimited, tabled here for readability): Note this is a limited version for the first implemetation with the most commonly used extension types we will add others once this is working O.K.

term_ID	term_name	Allowed_extension	domain	text to plug in for tool
GO:0004672	protein kinase activity	has_substrate (use substrate on page)	protein coding gene product	kinase substrate(s)
GO:0004672	protein kinase activity	happens_during (use during on page)	GO_process	kinase substrate(s)
GO:0034613	cellular protein localization	localizes	protein coding gene product	localized protein(s)
GO:0034613	cellular protein localization	directly_localizes	protein coding gene product	localized protein(s)
GO:0034613	cellular protein localization	indirectly_localizes	protein coding gene product	localized protein(s)

Many FYPO terms allow 'assayed_using' extensions. Do we need to list them all here?

Note The: extension will apply to all children (part of/is_a? need to check…)

On the gene page

1. if a term is allowed an extension, present a new page (or include in the annotation transfer page?)

From Midori: For transferring annotations, we'll have to add to the text to say extensions will not be transferred by default. Do we want to add the option to transfer extensions, e.g. by ticking an additional box? (I don't have a strong preference either way, but we should consider the possibility.)

We should also keep the ability to add extensions later (i.e. what it does now if you're logged in). People will need to be able to add extensions to transferred annotations, and just go back and add to any annotation if they forget or want to postpone.

From Val: May therefore need to be slightly different on the "phenotype" part as I don't know if we will allow annotation transfer in the same way here (because most of the annotation allele/descripion/cndition) won't apply it will be just as easy to change gene....

response from midori: We should still allow annotation transfer, but we could leave out the extension transfer part. Maybe show the allele description pulldown & text box instead? Should also think about whether we could show alleles whose descriptions have already been entered ... Off on a tangent: I've found myself wishing I could "transfer" an allele description to another phenotype term quite a bit. Not a priority before initial release, but something to think about for canto 1.1 (or 2.0 or whatever).

Text: "You have annotated $gene to $term. Can you specify the $extension [or $extension] based on the experiments in this paper?"

Q. How to present if multiple options? multiple rows? (check boxes? better option)

(has_substate,during) as opposed to (has_substrate|has_substrate)

midori: maybe one check box per extension, whether that extension is simple or compound? e.g.

 [ ] substrate geneA   [ ] substrate geneB  [ ] substrate geneC, during mitosis

2. Based on selection either i) exit, or ii) present a view with selected relation(s) and

a) auto-complete text field for domain to be typed if ontology b) gene products, I think it would be easier for users if we offer the gene list tickboxes as for annotation transfer (val: or a pulldown with the list?) (midori: tickboxes might be a bit more intuitive for picking more than one ext to transfer, but could get unwieldy with lots of exts ... on the other hand, with lots it's probably going to look mad in either boxes or a pulldown)

3. Antonia's suggestion for transfer (needs formatting)

You can annotate other genes from your list with the same term and evidence by selecting genes below:

[  ] geneX
[  ] geneY

you see this:

You can annotate other genes from your list with the same term and evidence by selecting genes below.
Note that annotation extensions will not be transferred by default; to also transfer the extension, please tick the corresponding box below: 

  [ ] geneX     extensions: [ ] substrate geneA   [ ] substrate geneB  [ ] substrate geneC, during mitosis
  [ ] geneY     extensions: [ ] substrate geneA   [ ] substrate geneD

4. Add a consistency check for out of date relations