WBConfCall 2018.05.17-Agenda and Minutes

Agenda

HelpDesk - Kimberly

Helpdesk Issues Still Open

• I am trying to have access to fosmid ...
• I'm only filling this out because your "need help ...

WBStrain ids

• would be good to get a list of people who need to be able to create new strains
• will reside temporary in the old NameServer at Sanger and then move to the new Datomic one on AWS

Minutes

• On call: Adam, Cecilia, Chris, Daniela, Juancarlos, Karen, Kimberly, Raymond, Scott, Sibyl, Todd, Valerio, Wen, Kevin, Paul D, Gary, Matt, Paul K

HelpDesk - Kimberly

Helpdesk Issues Still Open

• I am trying to have access to fosmid ...
  • On staging, we will begin providing the end sequences
  • We've never had the full sequence, so if we were providing a "fosmid" sequence, it might have been the intervening genomic sequence
  • What would users want?
  • It seems we could provide the intervening sequence but we need to communicate with users about what they're actually getting, i.e. a prediction.
  • Would this be a potential source of confusion to users, though, who may need to validate the sequence against a clone they order from Don Moerman's group?
  • Resolution: for next release, provide intervening genomic sequence but make it clear that it's predicted from genomic.
  • Chris will check and close the original ticket, if the changes Adam pushed are okay.

• I'm only filling this out because your "need help ...
  • We can't really address every possible browser add-on issue
  • Sibyl will respond and close.

WBStrainIDs

• Bulk assigning strain IDs to existing strains is easy.
• How do we assign new strain IDs at the point of curation?
• Will treat similarly to variations, i.e. use a name service to check if an ID already exists; if not, can request a new strain ID to use in postgres curation database.
• Who needs access to the name server for this?
  • Chris, Daniela, Karen, Ranjana
• Discussed institutional vs personal sign-on; personal is better, allows for assignment of WBPerson IDs
• What additional information will be solicited when requesting a new strain ID?
  • Public name is most important
  • Reference?
  • Genotype?
  • This really is an issue of what gets curated where.
  • Ideally, we'd have a central curation database (just saying) that would allow immediate entry of the data.
  • Another option is that all strain information would come from Caltech curation.
  • Would there be anything left for Hinxton to add or do wrt strain curation?
  • One caveat: if author's potentially correct data in the paper when they submit the strain data to the CGC, how would we reconcile that?
  • It would be good to have a shared data model with the CGC, maybe even share the same curation tool.
• Circling back, do we really need strain IDs?
• Are strain names unique?
  • Karen has found some strain and clone names that are the same.
• Todd did some analysis several years ago to see how many classes have overlapping entity names.
  • There were some wrt CDS and transcript; cell and cell group (no longer an issue), but overall the issue was not that bad.
• For AGR, Sibyl created an ID resolver page. Kevin tested and, so far, hasn't found any issues.
• There is also the issue of assigning unique IDs to genotypes. Should this issue be coordinated with assigning strain IDs?
• Raymond: can this change be coordinated with moving to the centralized curation database?
• Is use of the strain server an efficient way to curate?
  • Would it be possible to create a strain in the OA during curation? Would this be more efficient?
  • Strain IDs could be created during curation, added to postgres, and immediately available for curation, similarly to variation IDs.
  • If there were conflicts with the information in the CGC, we would ask the CGC to correct it, so that the information in WB and the CGC is the same.