WormBase-Caltech Weekly Calls October 2017

From WormBaseWiki
Jump to navigationJump to search

October 5, 2017

NAR Paper

  • Raymond will send around reviews

Mary Ann's transition

  • Cecilia will take over laboratory curation
  • Hinxton (Paul Davis) will take over nomenclature work
  • Hinxton will take over strain info, Mitani (NBP) alleles, sequence features (large scale)
  • Sequence feature curation (literature-based, small-scale) will move to Caltech
  • Molecule curation may be split; Hinxton could coordinate with ChEBI; GO-CAM modeling molecules handled at Caltech?
  • SOP? Karen wrote Wiki pages for molecule curation; she will look them over and can review with others
  • Metabolomics was recent big push (biggest effort needed); need to discuss with Michael Witting

AGR Press Release

  • Discussed briefly at AGR all-hands call yesterday
  • No one on the call had much info
  • Ranjana can post on WB blog on Oct 20th (or earlier?)
  • Will ask Stacia

AGR momentum

  • AGR disease working group
    • Lots going on, a lot of progress, still have much to discuss
    • Next need to work out how to pull in new data objects like alleles, strains, genotypes, etc.
    • Will need to establish a basic, consensus AGR data model for these data types/objects
    • How is AGR disease data display compared to WB/MODs? WB is generally (as are other MODs) ahead of AGR
  • Site-visit
    • PIs go to D.C. to discuss AGR
  • Wen working on lab meeting presentations; will present AGR progress, give tutorial
  • Raymond working on methods paper for SObA display
  • Raymond will start working on neural functional units with phenotype

Sequence feature curation

  • Really need to encourage authors to submit tables with specific sequence feature info
  • Need to make the data submission process easy, user-friendly
  • Karen looking into it
  • Work study jobs for students to curate? May not be preferred job for students
  • Temp jobs may be the way to go; but learning curve needs to be very short

WB Person curation/outreach

  • Cecilia contacts 1-6 people per week by email
  • She can ask for contact info: address, email address, phone #, etc.
  • How many persons do we have email addresses for?
  • Would be good to establish a 2-way communication with WBPersons
  • Should some people be demoted if they don't respond to emails? If they are not worm people?
  • We provide options to hide email addresses from WB person page
  • What do other MODs do? How extensive is the outreach to the community?
  • Cecilia reaches out when affiliations change
  • Some journals require use of ORCID IDs; makes identifying authors less ambiguous
  • Can ask people to update their intellectual lineage (can show lineage graph)
  • Juancarlos will query to see how many people with an email address don't have lineage info
  • One email from WB per month is OK
  • Can use Twitter more, when we do site visits


October 12, 2017

Strain curation

  • Paul Davis will take over strain curation
  • Many strains important to disease curation may not have been submitted to CGC, and may not exist in WB
  • Ranjana needs a way to create new strains for disease models
  • Ranjana will talk to Paul, and will add documentation about the issue

Phenotype data display proposal

  • Chris has been working on a new way to display phenotype data
  • This will largely depend on the new ?Phenotype_experiment model
  • Will hopefully solve a number of issues regarding how phenotype data is currently displayed
  • Mockups here: https://docs.google.com/presentation/d/1XvMN16B7RU2yPwjD5p9_TywMj0itln-djEWyGx2IxiQ/edit?usp=sharing
  • Goals are:
    • Clarify phenotype experiment meta data and provenance
    • Indicate when multiple perturbations may be responsible for a phenotype
    • Clarify when a non-coding gene is annotated to a phenotype whether the phenotype may actually result from disruption of a protein coding gene
    • Clarify when a phenotype results from a mutation that affects the exon(s) of one gene and only an intron of another
    • Clarify when a phenotype is a double mutant (or other complex perturbation) phenotype

How we handle/display multiple supporting interactions

  • As pointed out by Jae, some interactions appear very strange in the Interactions widget Cytoscape view when there are very many "independent" interactions/observations for a pair of genes
  • Example, tax-6 and rcan-1: http://www.wormbase.org/species/c_elegans/gene/WBGene00006527#0-8-10
  • We've tried to represent the "confidence" of interactions with thicker edges on the graph, but maybe this needs revisiting
  • Should there be a log-based increase in edge thickness, instead of a linear one?
  • Maybe evidence from the same paper should be considered as one observation?

Collecting person info

  • Could we request personal info when we ask for other info, like author first pass, community curation
  • Cecilia can manage feedback/input, and triage/defer issues to relevant curators etc.
  • Ask people to confirm current institution, lineage information
  • Probably need to keep the request minimal (a single line), so as not to overburden people
  • We need to point people to person update form
  • We could also do more through the website


October 19, 2017

Site visits, area meetings

  • About a 12-15 area and topic meetings next year
  • Chris will present at Baltimore Worm Meeting on March 16, 2018
  • Many meetings are annual (Topic, Seattle, New York), some bimonthly
  • Many of these arrange speakers well ahead of time, 6 months ahead for example
  • We should reach out to organizers soon, especially for annual meetings
  • We can combine WB, Textpresso, and micropublications
  • We can/should customize our talks to research areas of each audience
  • Many of the talks are about 40 minutes to 1 hour

Micropublication forms vs. WB data submission forms

  • Managing discrepancies still need to be worked out
  • Looking at differences of output formats
  • Customizable forms? Forms built by curator?
  • Karen: aside from paper details/meta data, forms can be the same

Phenotype data display proposal

  • Chris has been working on a new way to display phenotype data
  • This will largely depend on the new ?Phenotype_experiment model
  • Will hopefully solve a number of issues regarding how phenotype data is currently displayed
  • Mockups here: https://docs.google.com/presentation/d/1XvMN16B7RU2yPwjD5p9_TywMj0itln-djEWyGx2IxiQ/edit?usp=sharing
  • Goals are:
    • Clarify phenotype experiment meta data and provenance
    • Indicate when multiple perturbations may be responsible for a phenotype
    • Clarify when a non-coding gene is annotated to a phenotype whether the phenotype may actually result from disruption of a protein coding gene
    • Clarify when a phenotype results from a mutation that affects the exon(s) of one gene and only an intron of another
    • Clarify when a phenotype is a double mutant (or other complex perturbation) phenotype


October 26, 2017

GO CAM

  • GO CAM models do not go into ACEDB, yet
  • Kimberly is working on it
  • First, Berkeley group is working on producing correct GPAD files as output
  • Work is ongoing for a Cytoscape view of pathways
  • Noctua calls on 2nd and 4th Wednesdays of the month; anyone can sit in (ask Kimberly)
  • Noctua call minutes on Wiki (Kimberly will send around link)

Site visits

  • Wen setup Wiki page: http://wiki.wormbase.org/index.php/Meetings#Upcoming_Meetings
  • What's the strategy overall? Are we trying to make all meetings?
  • About 15 meetings to go to; would be good if we could make all of these, if possible
  • Kimberly will go to New York meeting
  • Chris will go to Baltimore meeting
  • Boulder, CO meeting; interested in having WB talk; sometime in May
  • San Diego meeting in January (Wen out of town); volunteers?
  • Chris could do Boston and Worcester meetings

Strains

  • Are strains in the nameserver? No, and Hinxton is likely retiring the nameserver soon
  • Can curators get access to a tool to request a strain object?
  • Juancarlos and Ranjana are working on tool (cronjob) that will automatically transfer the strain in the 'Requested Strain' field in the disease OA to the 'Strain' field once Paul Davis approves the strain.
  • Once a strain is entered in the 'Requested strain' field, it also shows up in the the new_objects.cgi form, for Paul D. to approve or not, but he still needs to be alerted to this.

Changes to phenotype form

  • We need a mechanism to clarify if multiple perturbations entered in the phenotype form represent multiple single experiments or a single complex genotype experiment
  • Will give users the option to indicate which of these two scenarios is the case
  • We discussed ways of managing this in a previous WB CIT meeting
  • Produced mockups of proposal: https://docs.google.com/presentation/d/1HFVn1anbdCu8pHJW0g27a5AvYLFVMyBy4nnrEdA1wKQ/edit?usp=sharing
  • Now will ask users for "RNAi target gene" along with "RNAi Reagent" and "Species" for RNAi experiments
    • We decided not to make "RNAi Reagent" a mandatory field
  • Optional fields
    • If there are multiple perturbations entered, we will gray-out (and make non-operable) the "Inheritance Pattern" and "Mutation Effect" fields as it will be ambiguous as to which perturbation each attribute refers to
  • With regards to the ?Phenotype_experiment model (and related model changes), we may want to store the inheritance pattern, mutation effect, and temperature sensitivity within the ?Phenotype_experiment object, as opposed to the ?Variation object itself, as these attributes are phenotype dependent
    • When authors report that an allele is dominant, or null, or heat sensitive without reporting a particular phenotype, a ?Phenotype_experiment object can be created that refers only to the "Variant" root-node phenotype
    • Then, when we want to display these attributes on an allele/variation page on WB, we can dig into the variation object's referenced ?Phenotype_experiment objects to pull out the attributes, but only from simple phenotype experiments in which we can unambiguously assign the attribute to a single allele
    • With this model, we won't need to place these attributes directly into the ?Variation model (or ?Transgene, ?Strain, or ?Rearrangement models)