Add new MF-MF relations to graph editor for creating example models for documentation

[vanaukenk]

In order to write curator documentation for using the new MF-MF relations, we'd like to add them to the graph editor pop-up window. The relations will also be added to the pathway editor with a separate ticket in that repo (coming).

For ease of display, I suggest that we not try to recapitulate all the new relations parentage in the upper section of the graph editor pop-up window, but @ukemi feel free to comment here if you think otherwise.

Here's a suggested display (only including direct parents and siblings here):

negatively regulates (RO:0002212) -directly negatively regulates (RO:0002630) -indirectly negatively regulates (RO:0002409)

positively regulates (RO:0002213) -directly positively regulates (RO:0002629) -indirectly positively regulates (RO:0002407)

causally upstream of, positive effect (RO:0002304) -constitutively upstream of (RO:0012009)

causally upstream of, negative effect (RO:0002305) -removes input for (RO:0012010)

Note that 'indirectly negatively regulates' (RO:0002409) and 'indirectly positively regulates' (RO:0002407) currently exist in the pop-up window but they aren't included in the upper section with the more commonly used GO-CAM relations and they have their old labels, 'indirectly inhibits' and 'indirectly activates', respectively.

Also note that 'directly inhibits' (RO:0002408) and 'directly activates' (RO:0002406), which are also both in the full list of relations in the pop-up window, are now obsolete.

Thank you.

[vanaukenk]

@kltm

The new relations have been added to the graph editor pop-up window on noctua-dev, but the label for two of the relations reflects an alternative label, not the main label currently listed in the latest RO release.

In the screenshot above, the labels for:

RO:0002407 (indirectly positively regulates) uses an alternative label, (indirectly activates)

and

RO:0002409 (indirectly negatively regulates) uses an alternative label, (indirectly inhibits).

[kltm]

Apparently there is something I'm missing. Going back to the source, minerva is sending this back:

[...]
"property":"RO:0012009","property-label":"constitutively upstream of"...
"property":"RO:0002409","property-label":"indirectly negatively regulates"...
"property":"RO:0002407","property-label":"indirectly positively regulates"...
[...]

These are the correct labels for the three listed relations, so not minerva. The next step where there might be an issue is in the ancient legacy "context" file, where things like color and edge glyphs are decided for the graph editor. Looking at that file though, none of these three are represented at all; I'm not sure where this is coming from... Going to do a little more digging.

---

I did another complete rebuild and restart of the system, including destroying all journals, and mystery labels persist. Hm.

[kltm]

@vanaukenk The good news is that I believe I've traced the problem back to the source; the bad news it that the fix isn't trivial and requires messing with some annoying old software and configurations. The fix should not take too long, but I wanted to give you the option to go ahead with the bad labels for these and I can follow up with the label fix after that. How much pressure is there here for the final fix?

[vanaukenk]

@kltm We do want the correct labels here, since we need to create GO-CAMs to use as examples for the new relations and I think it would be confusing to show models with different relation names than what we're trying to document. Let me know if you want to touch base about the schedule for fixing this. Thx.

[kltm]

@vanaukenk Yes, maybe I didn't explain myself well here. Let's try and touch bases a little. I'll ping you on another channel.

[kltm]

Noting https://github.com/geneontology/noctua/issues/804

[vanaukenk]

Tested on production and all appears to be working as expected.

Thanks @kltm