Version 2 (modified by val_wood, 9 years ago) (diff)


Planning Meeting Tuesday 11 July 2012 - Minutes

Agenda for this meeting

Tuesday 11 July, 16:00, UCL

Present: Val, Midori, Jürg, Antonia, Mark, Dan, Paul, Sam (part of meeting)
Remote: Kim
Apologies: Steve
Chair: Val
Minutes: Midori

Previous Action Items

  • Henning/using OLS
    • PK has had a chat; HH receptive but they don't have resources to allocate so PomBase would have to fund
    • will keep on back burner and revisit
  • Replication origins/Conrad
    • in progress; browser discussion (below) relevant
  • Selenium testing; server
    • London move & PHP upgrade have sorted speed issue
  • Link back to genomic regions; Mark to provide examples
    • What is this? BLAST output? linking to specific regions with description difficult ...
    • all BLAST issues on Jira, so no need for extensive discussion
  • Linking between browser and gene page
    • All in Jira, and all done
  • Interaction between EBI External Services and and PomBase
    • Plan is still the same: Eugene's team will take over managing git repository; we (i.e. Mark) can submit to repository and restart server in London
    • estimated to happen around end of September
  • Explain differences in GO totals (later)
    • documentation task for curators; in progress


Usage stats

Hosting data on tracks

  • Data must be on DAS server or incorp by ensembl to be available to public
    • i.e. no other way to show privately held data to everyone
  • BAM files - if user submits a BAM file to ENA, PomBase? can serve the data
    • Dan et al will help people with bam sub one-on-one if necessary
  • ENA will also archive KCF
  • WIG and BIGWIG
    • WIG suitable only for small datasets
    • ENA has not yet committed to archiving BIGWIG files, but PK imagines it should be fine with them

Future role of dev site

  • Mark and curators do all edits and updates on dev, then sync with/copy to www (i.e. no more having to make changes in parallel on both)
  • Envision these db/web instances
    • Mark's own box, for initial development; not visible anywhere else
    • proto-production ("staging" instance), which is then mirrored
    • split staging and live later if need arises
    • can make demo URLs available on ad hoc basis, e.g. for soliciting community feedback on new data or features

Curating strain data

  • plan so far: offer option -- not mandatory -- to provide background strain details, e.g. marker alleles
  • brief summary of strain/genotype-allele/gene/etc. relations
  • AL: curating strain details can add a lot of time to curation (depending on paper)
  • great utility generally acknowledged, but any more comprehensive approach would require us to seek funding for new aspect of project
  • curation interface complicated, challenging, and still under discussion
    • current thinking is to collect strain details in early step of curation
  • Sam's strong opinion: worthwhile only to show complete genotypes; partial genotypes much less useful
    • specifically, need marker alleles included
  • have not yet determined what we'll do with strain background details once collected
  • further discussion of issues
  • brief demo of allele curation interface, with discussion
  • see whether we can get a dump of data (IDs, genotypes, etc.) from strain collection in Japan
  • tangent: it would be useful to display alleles, e.g. on protein feature map; could also provide alignment of different alleles of a gene

Idle small talk, e.g. the recent misery of the weather

  • NZ winter is apparently cooler, but otherwise nicer, than UK summer

Curation update

  • Emailed
  • Brief discussion

Web site status and next priorities

  • Top priority: fix things that stopped working upon copy to www/preview removal
    • example: recent problem syncing between London servers
  • further discussion of syncing etc., with aim to avoid problem recurrence
  • curators will double-check current behaviour

Update cycle process

  • pipeline tests have worked properly, but we should not see regression and should raise jira ticket with high priority
  • still need to try live - can use Chado version with curation tool annotations
  • Mark having issues with API update, so live test will happen after he's back from wedding holiday
  • GeneDB decommissioning tasks and quick fixes; other quick fixes
    • on Jira
  • Handle annotation extensions
    • Needs a bit more work
  • Host a couple of HTP datasets
    • will become high priority once GeneDB decommissioned
  • PK keen to get other Schizo genomes & comparative analyses in soon
    • Fine with all as long as it doesn't delay GeneDB decommissioning tasks
    • bit of discussion of what's in those datasets

Genome browser usability issues

  • Browser feedback from the EMBO course and any pending browser overhaul
    • See Feedback from Nick Rhind and Paco Antequera
      1. Ease of configuration and uploading datasets
      2. The display
      3. zooming and browsing are not controllable
      4. resolution of tracks/data
      5. saving configurations
  • PK acknowledges cryptic error message upon upload of "bad" file format; also unintuitive configuration interface
  • PK: vertebrate genome users haven't requested extensive changes
  • Mark could make interface changes for PomBase browser
  • Val: is this a good use of Mark's effort, when changes would be generally beneficial?
  • getting this feedback from pombe community because of extensive heterochromatin work
  • Jürg: aware of many communities that just don't use ensembl browser; they use Gbrowse or similar
  • Val, Midori - similar impressions from pombe and other communities: ensembl web browser simply not used
  • PK: suggests functional genomics viewer for chromatin-related data types; can demo at next meeting
    • additional description & discussion
  • Jürg: main issue isn't missing functionality; it's that the interface is too complicated
    • further discussion, with examples from Val
  • PK: EBI wants to "keep number of services manageable"; its role in PomBase? is to provide ensembl deployment to meet community need
    • but feedback to date is that genome browsing needs are not being met
  • Val: ensembl interface requests should go to ensembl core team (even if they get assigned to Mark)
    • further discussion of ticket fate & communication
    • Jürg: improvements will serve ensembl well

New action items

  • Action: Curators look into acquiring data from strain collection in Japan
  • Action: Curators to monitor web site for problems (ongoing)
  • Action: Functional genomics views demo at next meeting (EBI)
  • Action: Val to circulate non-redundant list of genome browser requests & feedback from community
  • Action: PK et al. to create Jira tracker for Ensembl genome browser issues; PK to triage tickets


  • Next meeting
    • Will be in August; doodle poll for specific date
    • Invite Bert (EBI) for FG db demo
    • Meet at EBI if more convenient for Bert/demo; otherwise could do at Cambridge