Manager Call 2016-01-20

From GO Wiki
Jump to navigation Jump to search

Attendees: Melanie, David, Kimberly, Huaiyu, Paola, Judy (briefly) [The GO grant is due Jan 25th, so attendance was limited]

Minutes: Paola

Agenda

Annotation Files Available for Download

GOC Meetings for 2016

  • I think there's some confusion amongst curators about when the next site-wide consortium meeting is and what the aims of the post-ISB Geneva meeting are.
  • Geneva agenda includes discussion of protein complex annotation. Any other agenda items yet?
  • Any thoughts on when and where there might be a meeting after Geneva in April?
  • The schedule for the Geneva GOC meeting says that on Saturday April 16, "Meeting will run from 9am to 1pm." and "Optional break out groups may follow.". To book flights, we'd need to know if we should stay for the Sat afternoon or not please.

Minutes

Annotation Files Available for Download

Melanie discussed this with Tony. Here are her notes:

"1. GOA vs MODs (eg Pombase)

- GOA doesn't import IEAs

- GOA needs to be able to map back to UniProt accessions

- if there are problems with the AEs they may be dropped (i.e., keeping the annotation but dropping the AE)

One other potential difference between GOA species-specific files and those provided by MODs is that GOA does some filtering of the annotations to reduce redundancy:

- parent/child filtering of IEA annotations

- if the same annotation comes from multiple sources / with multiple evidence codes, GOA retains the one that is consideedr to be of the highest quality (where "quality" is based on the evidence code and who supplied the annotation)

2. GOA vs GO files

Pipeline: The gene_association_goa_human file is checked into the GO SVN under geneontology.org/geneassociations/submissions. GO scripts are then run and they move the file under the /geneassociations/ directory. Those scripts do some filtering including syntax checking and removing annotations to those species for which we are not the authoritative source (e.g. mouse, rat, arabidopsis, zebrafish...)

GOA does not scrape automatically the GO SVN to get files, though most of the GOA external source files are retrieved from the /geneassociations/ directory (i.e., post-submission processing)"

Discussion follows among all attendees. The following suggestions emerged:

- Add documentation about this topic, both as a mini-tutorial well visible from the front website page, and as FAQ(s) (Related: several users commented via the GO survey that the carousel on the front page is too big and not very useful => we could make it smaller at least, then the top articles would be a lot more visible.) For example, we have an FAQ about finding annotations in QuickGO that can’t be found in AmiGO, but there is no FAQ on using AmiGO. Note that the ‘Help’ link on AmiGO points to the wiki, we should at least move that documentation to the website… Related: users who access GO information directly via AmiGO has no direct link to documentation or FAQs that are available on the website. That causes a loss of valuable information that we need to address. Huaiyu - who’s responsible for AmiGO documentation?

- AmiGO allows a maximum download of 10,000 annotations. We may want to speak to Seth to adjust that limit, be able to increase it manually like you can do in QuickGO.

- On the GO website, under downloads, we need to help users get to AmiGO too.

- It might be worthwhile to do more naive user testing for AmiGO.

All these items imply discussion on allocation of resources…

GOC Meetings for 2016

Judy: not sure if we’ll have another GOC meeting in 2016, let’s clarify at the next managers call.