Manager Call 2020-01-08
- 1 Agenda
- 2 Discussion points
- 3 Points from previous meetings
- Agenda: Pascale
- Minutes: Laurent-Philippe
- Present:Pascale, David, Judy, Seth, Laurent-Philippe, Paul, Kimberly, Huaiyu, Chris, Suzi
Looking back at the Alliance meeting
- Biological Function group status : ribbon mostly done; next was supposed to be GO term page with reusable widgets, including for GO-CAMs but currently on hold. May need to update the group charter. Also need to provide group description for the Alliance website
- We need someone from GO to monitor 'gene description' and 'interaction' groups (LP had good contacts with Chris Grove during F2F in Boston)
- Increasing desire within the Alliance to have more integrated User Experience (eg UIs allow only for separate access of data eg GO or disease or expression)
- Alliance needs to reprioritize some projects and really focus on priorities. All agree that reusable widget components will be the way to go. The GO term page may be more a GO delivery than an Alliance one.
- GO data comes for the ribbon comes from GO but GO data for gene descriptions currently comes form the Alliance. There are some sync issue there but may not be critical, especially considering the gene descriptions comes from the integration of several data.
- Need further discussion for an Alliance representative that would be more biologist than technical. Paul proposal: instead of participating to all the calls, maybe just brainstorming with some groups at some point in time. Discuss with Valerio (head of gene descriptions) every 3 months for instance to keep track of progresses and issues.
Status of Projects
Priorities spreadsheet: https://docs.google.com/spreadsheets/d/1lZO0cf64_eufimUhEdcpiZMGEmvF6_7BW-GQWBXR3Mw/edit#gid=0
Folder in Google drive with specification documents: https://drive.google.com/drive/folders/12iA4Y8F0Acz4TaHyoo2mk-vWJ0cEl_sD
- 3.1 Migration of MGI and WormBase annotations to Noctua, including 'round tripping' to make sure imports/exports are working
- 4.1 GO-CAM ShEx specification v1.0 completion: Close
- 4.5 Curation history in Noctua and GO-CAM
- Set up a call where we discuss specs and implementation strategy: Seth, Kimberly, David, Chris, Paul, Pascale. etc ?
- 8.7 Improve GO API & support the GO-CAM and documentation
- Chris: status update - must do the requirements
- 14.2 golr indexing of go-cams (also required for 8.7) (Alliance & alliance-required developments)
- Chris: added as product owner - must do the requirements
Outreach: monthly seminars
From Chris: What I propose is that we schedule one-off zoom calls with select groups such as these, where we answer questions, educate, solicit feedback on how they want to obtain the information, share ideas on algorithms. The notes from these can be used to seed further technical documentation.
I would say one call a month, duration one hour, with up to one hour for prep (no ppt, keep it informal).
Key technical people would be: Laurent-Philippe, Ben, Jim, Seth. We may want to have Kimberly or David available for biological questions, but for more technical calls this may not always be necessary (trying to streamline and be efficient here).
Obviously we have to prioritize this, but we need to lay some groundwork for the next grant, take advantage of any current interest and momentum. The structure I propose here should be lightweight and not too onerous. It may end up driving requirements that would then need to be prioritized.
Action: Chris starting with Dexter and Peter Robinson to setup meetings
Status. Draft is here. Validate overall schedule: lunch time issue on Tuesday and Thursday: move user meeting at the end ? Registration fees: waiting for USC feedback
Schedule validated; Friday remove for the moment and open discussions with local teams can be done on Monday
Points from previous meetings
Implementing a testing process
- We have had some unexpected side effect of the deployment of some code, last example was in Noctua.
As we move towards being the provider of the annotation tool for various groups, we need a way to avoid these situations, by having a formal process of testing and releasing updates to the code.
More documentation on software architecture
Programmers and Product Owners (probably except for Seth and Chris) would really benefit from having a software architecture diagram - does that already exist ? - Specifically: how do the Biolink API, the GO Triple Store, Minerva, the GO API, GOlr, etc ? relate to each other -
- what is the data flow?
- which ones have dev/prod versions?
- how are these various tools incorporated in outward-facing products?
- .... ?
- we need to know what is implemented - not just the final goal
Zenodo bug during the last release
- What was the problem in the end ? Why was Zenodo not being responsive?
- Is it is a good idea to rely on external resources for this ? CERN is a research project - do they really guarantee to be stable for 'ever'? what is our backup plan, should they not be able to support maintaining our archive ?