Add GO-CAMs to NDEx

[kltm]

Project link

https://github.com/orgs/geneontology/projects/157

Project description

This project aims at adding GO-CAMs to NDEx (https://home.ndexbio.org/about-ndex/) for:

• increasing community access and awareness of GO-CAM models
• basic "enrichment"

PI

Chris

Product owner (PO)

Chris

Technical lead (TL)

Patrick

Other personnel (OP)

N/A (Possible input from Sierra and Dustin)

Technical specs

https://docs.google.com/document/d/1g0gVCJ9FeB1Yvlw218t3WYSCg1ou2GqZXzwMY54mrkk/edit#heading=h.gx5r3bwu5gee

Other comments

N/A

[kltm]

@pgaudet At Chris's request, this would be put into "Miscellaneous".

I'm not sure who is working on this yet.

[kltm]

@cmungall Would it be possible to fill in a bit more description here? I'm assuming that Sierra or Dustin is who you're thinking about here?

[kltm]

Also see #56

[kltm]

GO-CAM tech demo for how users can helped with this work

[kltm]

@pgaudet I have a note that says that I am to tag you onto the existence of this project.

[kltm]

Now looping in Patrick, as AmiGO fix is wrapped and this is next on the list from the meeting.

[pkalita-lbl]

Summary of current progress:

• Code has been developed in https://github.com/geneontology/gocam-py which is capable of converting a GO-CAM into CX2 format, which is uploadable to NDEx.
• I created a GO group on NDEx, uploaded about 20 test networks converted from GO-CAMs, and shared them with the group. The uploaded networks are still private as we continue testing.
• On 2024-08-20, Chris, Seth and I met with folks from NDEx to discuss various aspects of organizing networks in NDEx; keeping NDEx networks in sync as the underlying GO-CAMs are added, modified, and removed; and the technical requirements around having the GO-CAM networks indexed by their IQuery tool. Meeting notes here.

Next steps:

• The NDEx team recommended creating and using a GO user in NDEx for managing networks, as opposed to a group. I will create the user and ensure Seth and Chris also get access to the user account. Existing networks will get transferred to the new user.
• I will also share our existing test networks with members of the NDEx team so that they can take a look at them and provide general feedback.
• I will work on updating the GO-CAM-to-CX2 translation code so that the resulting CX2 networks adhere to the specification required by IQuery (will create an issue for this work).

[kltm]

Talking to @pkalita-lbl :

---

Where we’re at right now is:

1. We met with the NDEx folks and they gave us a good overview of how IQuery work and how the networks we upload need to be structured to be indexable (you already know that; you were there)
2. I made issues to capture the changes needed on our end: a. https://github.com/geneontology/gocam-py/issues/7 (done) b. https://github.com/geneontology/gocam-py/issues/8 (in progress; I need to talk to Chris about this at our next 1:1) c. https://github.com/geneontology/gocam-py/issues/6 (not started yet, dependent on the other two)
3. Once those three issues are done I’ll do a one-off convert-and-upload of a bunch (all? all human? TBD) of GO-CAM networks and ask NDEx to try indexing them. This may result in a bit of a feedback loop.

[kltm]

From @pkalita-lbl

Updates from this morning:

• Rudi had good feedback for us and I’ve captured it [here][https://docs.google.com/document/d/1krMV2lkw-i_ZH0EUEMQ_chgDYLVDn2zR2F8BASVvovg/edit#heading=h.s0igijnovzoz]. I told him that we’ll try to work incrementally on that.
• He said that they should be able to start doing indexing of the GO-CAM networks on one of their test IQuery servers. I got the impression that the test server would still be publicly visible. Seth stressed that we’d love to have that by Oct 15 and Rudi didn’t seem to think that would be a problem.
• I just got an email from Rudi saying “Your GO-CAM networks have exposed a bug in NDEx” :grin: He had suggestions for how we can update our networks to work around the bug while they make a fix on their end. I’ll try to get to that tomorrow at the latest.