Add ability to filter redundant annotations

[cmungall]

• Add a flag to the golr schema for annotation documents to state when an annotation is redundant
• Populate this in the golr loading code
• Add a test case
• Add a 'include redundant' toggle facet view; default is off (i.e. we remove redundant by default)

TBD:
do we want a globally redundant flag? Or make this relative to evidence code?

See also:
See also: http://jira.geneontology.org/browse/GO-144

Priority: post-Bar Harbor

[kltm]

Just to clarify, we're talking about a new field that is just binary? A la "is_obsolete"?. If so, I'd rather get it in sooner rather than later (even if we don't load it) since I'll be doing another schema update process soon for the complex annotations.

[kltm]

After conversation: add "is_redundant_for" to annotation schema.
It will be aggregated out during main GAF loader (a la the other aggregate features).
Would take various string values, or nothing at all in the case of completely non-redundant annotations (optional).

[kltm]

Added to schema, but not loaded yet.
Currently single-valued.

[cmungall]

Reminder note to self: for bioentity docs we should always filter redundant annotations in the individual lists

[kltm]

If I'm following you, I'm not sure how this would work without a separate associated term list for every type of redundancy.

[cmungall]

This would only be for the most general kind of redundancy

On Wed, Sep 25, 2013 at 10:46 AM, kltm notifications@github.com wrote:

If I'm following you, I'm not sure how this would work without a separate
associated term list for every type of redundancy.

—
Reply to this email directly or view it on GitHubhttps://github.com/kltm/amigo/issues/43#issuecomment-25108919
.

[cmungall]

copying comments from @ValWood:

+1
At PomBase we do this with IEAs NAS and TAS,. We only use these to supplement manual annotation and discard anything which is less specific than any manual annotation. Nobody misses them and we remove 87% of our IEAs in this way (all but ~4000 currently).

This has lots of other advantages as well making the annotation easier for our users to consume, the bad IEA annotations are now easy to spot.
https://github.com/geneontology/go-annotation/issues

http://www.slideshare.net/ValerieWood/pombase-internal-rules-for-curation-using-ontologies
slides 23 -28

[pgaudet]

@cmungall Can we have the decision parameters to flag an annotation as redundant ?

[pgaudet]

Redundant with geneontology/go-site#943