WBConfCall 2016.04.21-Agenda and Minutes

From WormBaseWiki
Jump to navigationJump to search

Agenda

1. Request to add WormBase to list of SciData repositories . See git https://github.com/WormBase/website/issues/4753

  • Follow up on Biocurator meeting.
  • Daniela says they work with repositories (e.g. zfin) but do not integrate data into the repository, did they ask how to integrate data ?
  • Open access, no money involved.
  • Karen says it's about little details of research along the lines of Worm Breeder's Gazette.
  • We should give them our details, we don't gain anything by not being on their list and all the other MODs are there.
  • Mary Ann will answer the questions, will email specific people about specific question.


2. Request for tissue enrichment analysis reverse proxy. See github issue https://github.com/WormBase/website/issues/4750

  • Todd set up the proxy.
  • Raymond will look into the exposed URL from forwarding.
  • Spell will also need forwarding.


3. Biocuration Meeting round up (Mary Ann) nb Kimberly is on leave.

  • Link that I will explain: https://docs.google.com/document/d/1v7nJU1d8vRtnpqKfYDmDjeQvIHIcpJx1kUlbYxvZH6k/edit#
  • Mary Ann and Kimberly went to Biocurator meeting. Geneva was lovely.
  • Should be automatic for authors to submit data to database without being forced to by journals.
  • Curators can go to labs to show users how to do it.
  • Talked about collaboration with journals.
  • Workshop about what we can do now to promote community curation, not new developments.
  • Phenotype response from users at 18%, steady user submissions.


4. update on uMOD (pws).

  • b. Any news on NIH funding for other MODs? no
  • Fish, mouse, and yeast know they're going to get funded, but don't know how much or how long.
  • uMOD moving along.
  • Yeast, worm, mouse, rat, gene ontology (missed any ?) discussing forming a consortium, deciding how deep.
  • Call tomorrow to organize more.
  • What are we doing for website visualization widgets.
  • Plan for unified website frontend, decide architecture over a year, but need proposal by July.
  • What tools do people want, and what do people have.
  • PIs meet in two weeks in Chicago airport.
  • WormBase people, reach out to counterparts.
  • Give everyone the same type of data other MODs do. Need to figure out a process to evaluate with different MODs what is done for specific data.
  • If we need something simpler, it will be in addition to what we already provide.
  • Chris emailed how interaction displays are different across MODs.
  • Karen is already talking with zfin people about curation comparisons, we will see about expanding these to include other MODs.
  • Will make email list or staff for all uMOD.
  • If there is no corresponding curator at a MOD, email the PI.
  • Keep going with Datomic, engage other MODs about technology choices.


5. Significant improvements to JBrowse:

  • Considerable speed improvements (loads in ~10 seconds versus 2 minutes or never)
  • Added a plugin to switch track selectors (either the pull out thing we've been using or a list of tracks with a checkbox on the left of the display). See this for example, where the button in the upper right corner will switch between the two.


6. We need globally unique IDs for WormBase entities.

  • ZFIN uses this format: three letter database- class-number, ex: ZDB-GENE-980526-388,ZDB-TSCRIPT-110325-1492, ZDB-MRPHLNO-110614-1, ZDB-FISH-150901-13768
  • Please see the following documents for the RRID project that lay out the use cases, end users, and pros and cons of different identifier formats:
  • Unique IDs at the class level is not enough for third parties to refer to WormBase.
  • Half a million entities like Transcripts-CDS, Strains-Cells overlapping.
  • At EBI DOIs get assigned to objects that don't already have identifiers.
  • DOIs are overkill, but we need a unique ID.
  • Tracking overhead for assigning random numeric IDs, less overhead to make existing IDs unique.
  • Strains cannot be unambigiously resolved as is, the names often are identical to clones
  • Datomic has universally unique IDs, gets added as a field.
  • Ask other MODs how they do it ? It hasn't been resolved yet. Todd can send examples.


Help Desk