Manager Call 2020-04-01

From GO Wiki
Jump to: navigation, search

Agenda

  • Agenda: Suzi
  • Minutes: Seth
  • Present: Pascale, Suzi, Seth, Kimberly, Judy, Chris, Laurent, PaulT
  • Regrets:

Discussion points

Community curation with post-docs/grad students

  • Any updates here?
    • May have an influx of C. elegans researchers creating GO-CAMs with an eye towards accompanying micropublications
    • Will require initial training and ongoing supervision (at least for awhile)
  • Want to follow the SOP for new user onboarding
  • Need a solution with as little training as possible, little effort from GO
    • SGD's "author response"= quick summary form (https://www.yeastgenome.org/submitData), accessible by curators on another site
    • Canto- heavier than SGD
    • Google form- lighter than SGD's setup
    • Targeted audience- viral focus?
  • GO CAMs on pre-pubs
  • Partnerships- may be easier for novices to enter through something like Pathways (https://www.yeastgenome.org/submitData), getting user to think about how to create this model, or if they've published schematics that can transfer to GO-CAMs
  • Also have a scheduled call with Xenbase curator(s) for Noctua training

GO virtual meeting Paris

  • New eventbrite registration is available

COVID-19

What can/should GO be doing, not just in terms of virus annotation, but wrt helping direct energy towards curation efforts, e.g. from graduate students who can't go into their labs?

  • Waiting for UniProt IDs to be integrated into Noctua
  • Are about small number of papers

Mockup for COVID-19 page: LP is contact, will email Managers list

  • Would like to see EA link on COVID-19 page
  • Be careful of nomenclature
  • Immune or host Slim may be more appropriate? Alliance is general, good for Human
  • Noctua annotation: Annotating UniProt protein chains - can we use the UniProt_PRO_chain as entities ?

File Formats

  • https://github.com/geneontology/go-annotation/issues/2864
  • GPAD/GPI 2.0 - need to decide on realistic timeline for switching over
    • How much lead time is going to be needed?
      • GOC, GOC members, users
        • WB - earliest production release would be late August/early September
      • Will ask for this feedback on the annotation call
  • Still have some outstanding questions wrt standardizing representation of IDs, e.g. proteins, that can map to more than one gene (software call?)
  • We are suggesting that groups that use UniProtKB accessions for these proteins work with UniProt to create gene-centric entries, but in the meantime we should give guidelines about how they will be represented in gpi (and then in neo)
  • For the DB_Object_Symbol and Parent_Object_ID fields, should we:
  • 1) Make cardinality of DB_Object_Symbol 1 or greater so that every relevant DB_Object_Symbol is in a pipe-separated list? That way, we won't have the situation like what we currently have at WB and UniProt, where we pick just one of many.
  • 2) Make cardinality of Parent_Object_ID 0 or greater so that we capture every relevant associated parent object, e.g. MOD gene IDs? It looks like we're doing this already in the current WB gpi, but the 2.0 specs say the cardinality here is 0 or 1.
  • GAF - proposed update wrt gp2term relations
    • Use term labels as is or modified with underscores?
    • Need to create a ticket for comment and also think about release timeline here

Discussion

Community curation with post-docs/grad students

Judy: worries about applying our own resources
Kimberly: PaulS has some people, tie them in with micro-publications; Kimberly may be a little diverted for this
Chris: deadlines may have to be shifted, of course
Kimberly: Now doing a Friday session with a Xenbase curator; David and Dmitry to sit in as well; have started onboarding them
PaulT/Pascale: It will help us get feedback and an interesting experiment

GO virtual meeting Paris

Laurent: not much news, but some new registrations; registration is still eventbrite

COVID-19

Chris: Weaken language?
PaulT: Already proven, no worries.
PaulT: Still a few bugs there.
Judy: Overall looking good.
Judy/Laurent: Article, blog?
Judy/PaulT: Write up overview of how to do a GO analysis. Have Laurent lead out on that.
Judy: Ten steps on using GO in data analysis? PLoS; very timely right now.

File Formats

Kimberly: timeline would be good.
Seth: The only technical requirement is imports
David: MGI is still doing “1.1”, or a partial import
Chris: if it looks like 1.1, we can look at a converter
Seth: let’s just land on 2.0 sooner and get the code all in one spot
Kimberly: will check in at WB,
Kimberly: We have been saying that we want this, but have never had a date.
Pascale: What is our policy here, if there is no technical reason?
Chris: No harm in using GAF.
Seth/Kimberly: So what about kill GAF? If GPAD isn’t “necessary”, except for imports, what do we do about GAFs?
Chris: gp2term relations should be discussed first as may enlighten discourse;
Paul: GPI for NEO for Noctua; GPAD for import for Noctua, better QC; what else?
Chris: It’s 2.2. A month warning, but would only be for internal input.
Pascale/Kimberly/: need a formal spec.
Chris: to prioritize, let’s see progress on gp2term
PaulT: provider-based relation selection for consortium members; thinks we should propose a deadline for an answer.
Kimberly/Pascale: P2G users can already distinguish.
Pascale: worth doing, announce on annotation call
Chris: 1) spec, 2) announce on call, 3) clean up a little

Publications

Kimberly: To check in with Val

Polyproteins and Noctua annotation

Chris: A little worried about IDs, was hoping that
Pascale: UniProt is planning on unmerging and making separate identifiers
Chris: Timeline?
Pascale: Timeline a little unclear, has been on their backburner for a while, but now there is interest...
Chris: This is preferred.
PaulT: What do we do in the interim? Happy to hand build a GPI, if necessary.
PaulT/Pascale: Likely just need to use the “temporary” IDs.
Chris: The ticket needs to be updated--two species not mentioned.
Pascale: What about GO-CAM models on humans? Might be more interesting than viral genes?


Punted items

MGI and WB Imports

Status updates

  • WB
    • Alex will work on preparing a test GPAD file with annotation history for WB - hopes to have it on the GOA ftp site soon
  • MGI
  • Have a periodic follow-up call to review progress?
  • Malcolm already attends Judy's meeting, TO-DO: invite him (and/or xenbase reps) to GO-annotation calls. Need to make sure XB maintains own annotations, is in line with GO SOPs.

Pathway Genome Databases and the Alliance

See Chris' email

  • Explore bringing in other existing curated pathways in a similar way that we did for Reactome
  • Follow-up to Alliance call with Peter K. two weeks ago?

Publications

  • Val's matrix paper is available as a draft
  • Authors have reviewed and made comments
  • Do the GO PIs (besides Chris who is last author) also need to read?