Manager Call 2019-09-04

From GO Wiki
Jump to navigation Jump to search
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Agenda

  • Agenda: David
  • Minutes: Laurent-Philippe

Review action items from last week

  • Seth: Set up a wiki page for GO users meeting, and start announcing. (DONE)
  • PIs/managers: Start on the GO meeting agenda (this is a reminder. In the Google drive https://docs.google.com/document/d/16ZCV1vyTHSEf20z_PO4F8S-uqD2ZvnyExV0McMWnFI0/edit)
    • Current projects
    • Ask if people want to present anything.
  • Kimberly: talk to Pankaj to see whether they can clean up their file, and correct annotations DONE - Pankaj is OK for us (GOC) updating Gramene annotations as needed; Planteome is working on a new pipelines for generating annotations. He will try to send someone to the GOC meeting.
    • We will ask if we can handle any annotation edits in P2GO for now.
    • TO DO: Need to make sure GOA is the source from Gramene. Check with Alex.

GO Meeting general

Agenda first draft (by Pascale) https://docs.google.com/document/d/16ZCV1vyTHSEf20z_PO4F8S-uqD2ZvnyExV0McMWnFI0/edit


Release QC

Pascale, David Kimberly https://docs.google.com/spreadsheets/d/1jrd-TiAoq7WuSPMQKryP-ci2wRJeq5ELIwVY6Hy3CDs/edit#gid=0

  • All looks good, except RNA Central has a lot of changes that we don't understand. Would it be possible for a specific source to go back to the last release version of their data, until we sort out the explanation for the large changes ?
  • Release dates inconsistency between GO and annotations - we should have a single date for our release, this is what we ask people to cite.
  • Should we announce major changes ?
    • ontology: new terms/obsolete/merged
    • annotations: New plant genomes in PAINT - 20 new plants, almost 900,000 new annotations

Review of project process

Anybody has anything to reports ? Especially blocking issues/limited resources for high priority projects


Collaborators Webpage

We need to decide who we put as collaborators and as contributing groups. There is some redundancy. https://github.com/geneontology/geneontology.github.io/issues/130

Need for a IT/Cloud strategy

While not yet critical / urgent, it is important to have a IT/Cloud strategy for the project to ease the sharing of resources, virtual machines, protocols, loggings etc. Laurent-Philippe created a document for the USC AWS account in April. It could serve as a basis for discussion. Example of ticket affected by the lack of IT/Cloud strategy: https://github.com/geneontology/operations/issues/30

Problem to solve: GO servers and resources are scattered across multiple IT and multiple AWS accounts which are neither listed nor shared. Aside from inconsistency, lack of transparency, being cost ineffective and problems to maintain the architecture in the long term, sharing resources both for development and production is complicated and susceptible to create non standard applications and pipelines. A shared IT/Cloud strategy would also allow developers to quickly share and launch any GO tool for deep and quick iterative testing.

Proposal: allocate some time for the next period (October-May 2020) to:

1. list all currently used IT & cloud resources (servers, storage, DNS, cloud accounts, etc)

2. define together a group strategy and SOP for those IT & cloud resources

3. define and share base images and docker containers to launch all GO tools for deep testing and rapid development


Examples of what this SOP could define:

  • how many on-site servers and AWS accounts we have and how they are organized to share resources, security permissions and costs (e.g. maybe we still want 1 LBNL and 1 USC account, but then both should be linked to a parent GO account giving permissions to each organization)
  • who are the current admins of these servers and AWS accounts
  • who to ask to access more restricted resources (e.g. Seth is the only one to know where the geneontology.org DNS is hosted and to have permission to alter any URL *.geneontology.org ; same thing for release.geneontology.org)
  • where and how is administered help@geneontology.org ?
  • how to create new AWS accounts for GO developers (e.g. Laurent-Philippe, Ben and Anushya had to use their own accounts, hence not sharing any standard, resource, images or security permissions)
  • what default security permissions are granted to GO developers (e.g. ability to launch a test server, to write in some locations, etc)
  • list all images and docker containers to launch and test in a standard and reproducible way any given GO tool (e.g. Noctua stack, GOLr, NEO, GO API, etc)

Cross references on GO website/AmiGO page

Present

Judy, David, Huaiyu, Seth, Suzi, Kimberly, Paul

Action items

  • ACTION ITEM: [Kimberly]: Open a ticket to see how we'll handle the Gramene/Planteome data 
  • ACTION ITEM: [To discuss later] Make sure that the Types are used correctly - right now for example C. elegans provides genes and PAINT uses proteins for the same type of identifier. This is misleading for users.
    • July suggests widening the definition of protein to include gene and protein.
  • ACTION ITEM: [Seth to open a ticket] Date on the ontology is different from the release date: http://skyhook.berkeleybop.org/release/ontology/go-basic.obo - in the go-basic.obo the date is Aug 31, the release is from Sept 1st.