KANBAN-124 GO Ribbon issue: duplication of some terms when "all annotations" is selected

[MagdalenaZZ]

Please see Alliance ticket https://agr-jira.atlassian.net/browse/SCRUM-1031 and https://agr-jira.atlassian.net/browse/KANBAN-124 for more details + screenshots.

[kltm]

After some discussion, we'll be doing a little research to see if this is a data issue (incoming payload is incorrect) or a ribbon drawing issue.
@tmushayahama

[kltm]

Who would be the best person to help @tmushayahama orient himself in with the UI so he can work on just this code?

[vanaukenk]

@tmushayahama
Note that the behavior I was describing on yesterday's software call is also described in the Alliance ticket:
https://agr-jira.atlassian.net/browse/KANBAN-124

Namely, the duplication is apparent when viewing 'all annotations' but is gone when clicking on a specific term box in the slim.

Contrast 'all annotations' on this page: https://www.alliancegenome.org/gene/HGNC:8896#function---go-annotations

With just selecting 'catalytic activity' from the MF slim.

The duplicated 'phosphoglycerate kinase activity' evidence seen under 'all annotations' disappears when just selecting 'catalytic activity'.

[vanaukenk]

Here are screenshots of what this looks like on the human PGK1 page.

The first shows the duplicated evidence for 'phosphoglycerate kinase activity' when the all annotations box is selected:

The second shows that the evidence is not duplicated when the box for the MF slim term 'catalytic activity' is selected:

[lpalbou]

It is an interesting issue.. I spent some time trying to figure out where this was coming from. It doesn't come from the ribbon but it comes from the GO API and more precisely the slimmer function.

The slimmer function is the piece of code used to retrieve all annotations under a term or aspect for a given gene. When clicking on "MF", it does show the correct annotations, but when clicking under "All", and as shown in the screenshot, the PK activity evidence appears duplicated. After checking the GO API request, one can see that we find PK activity under both a slim of BP 8150 and a slim of MF (3674):

PK activity under BP:

PK activity under MF:

The ribbon table group all annotations per term, and since this exists twice at the level of the GO API / slimmer, it displayed it twice. For reference, this is linked to an old issue we had, where when clicking on a BP term, you could see MF terms. I then created a filter to remove any MF terms that would be "slimmed" from a BP term. A similar strategy could be done when clicking "All", by removing any MF from the BP slim section. This is probably better be done at the GO API level (@sierra-moxon) but this could be intercepted and filtered out by the ribbon as well (@tmushayahama ).

@tmushayahama @sierra-moxon

Hope this helps

[kltm]

These have been marked as Done on the Alliance side. If that sounds right, go ahead and close.