Manager Call 2015-10-21
GO survey update
- Action items: post news article on GO website; forward to collaborators and MODs mailing lists if not done already
- AI: Moni will post
- AI: Moni and Chris will check why others can't post on drupal
- Received 137 responses already
- Sent to many mailing lists and individual addresses, including authors of GO papers as provided by Kimberly, GO helpdesk users, and people who took an EBI survey last year and declared they use GO
- UX people at EBI advised multiple collectors, to help us differentiate users
- Paola set up different collectors; e.g. we know we have 35 responses from Pombe community; 18 responses from Sanger; 8 from Textpresso
- We can still pool all responses, or we can differentiate
- We used to use Bluejeans, administered by Rama
- Can we continue to use that? If not, what are the other options?
- AI: Kimberley will contact Mike to get access for all go-managers, or is that too many people?
- We want at least 2 people
- We don't need for ontology calls, we use goto
- We need for other calls like lego
- Bluejeans generally worked very well for the calls
Is someone taking over as annotation manager?
- PaulT: will be talking about this later on go-top call
- AI for managers to report back
- PaulT: request for comments
- K: Rama was wearing hats for advocacy
- Chris: align what's in Trello and website
- PaulT: pairing up two people a good idea? One experienced one relatively new. K: this is a good idea.
- K: Rama was writing up a list of her responsibilities, gave to Mike?
Use of Accessions and Identifiers in Curation
- Following up on an item from the 2015-10-13 annotation conference call
- Typically, GO curators use UniProtKB accessions, MOD gene IDs, RNACentral IDs, and IntAct or PRO protein complex IDs in Column 2
- In Annotation Extensions (and also to some extent in the With/From column) curators may be opting to use other IDs to try to indicate a more specific entity, like an mRNA transcript
- GO agreed some time ago that reference protein accessions and gene IDs would effectively be interchangeable and would mean one or more entities associated with that accession or ID
- Are annotation extensions now being treated differently? Do we want that?
- Can we handle isoforms in annotation extensions?
- What is human readable/understandable vs machine readable wrt entity IDs in annotation extensions?
- The ncRNA annotation guidelines that Rachael has put together in a paper should address this issue clearly and that paper is nearly ready to be submitted
- Chris: we need to use the correct type in extensions
- David: what if we don't know the isoform for the transcript
- David: 3 different identifiers for genes
- K: Proposed decision tree
- David: Example rule: if you're talking about a gene always use MGI,RGD, etc
- K: we need specific isoforms
- Pascale: needs to take this to another discussion
- K: include Rachael
- AI: Discuss solution for LEGO on LEGO call (Monday)
- AI: Then we (K, cjm, David) will report back
Topics Leftover from 2015-10-07
File names and content (Melanie)
There was a proposal to update the content of the files the GOA team is producing, specifically the current goa_"species" file would disappear, and the current goa_ref would be renamed as goa_"species". [...] Minutes: Melanie to talk to Tony about only changing the GOA file names - if no issue go ahead, if issue organise call.
AI: Melanie will send an email to Tony, Chris and Paul to follow up on trello item "reference proteome should be default" https://trello.com/c/holv0gEG/151-reference-proteome-should-be-default
AI: will move the items below to next managers call:
Below not discussed:
- has_regulation_target can all annotation with this relation be updated to regulates_transcription_of? Yes. We have to figure out who will do it.
- Tony will send a list of annotations that have deprecated relations to all the groups.
Action item: groups will take time to rehouse the annotations (we will not delete/drop those annotations)
- contact a text miner to determine which papers (of the reactome set) correspond to which players (gene products) in the reactome annotations. Peter d’E will contact Jo McEntyre who may have already done this (cc-ing Claire).
Tracing the Annotation source
- We need to form a working group to figure this out.
- In 2014 we decided to annotate ChiP experiments to chromatin (CC) and not chromatin binding (MF). We will continue with that practice. We should update documentation on this.
- Follow-up: collaboration with Trey Ideker's group on autophagy (David)
- This would be really good to show community engagement. Status update, publication plans, etc.
- Moving tickets from Jira (EBI instance, NOT Jira helpdesk) to GitHub (David)