WBConfCall 2024.06.06-Agenda and Minutes

HelpDesk

• https://github.com/WormBase/website/issues/9524 homework ?

• Fwd: WormBase allele queries Did this go to help desk ? Didn't see email and don't see it on github

Agenda

• Town Hall and subscription funding
  • Paul S : Had two town halls. Good support for subscriptions and fees. Interest in subscriptions. Going to have three town halls. Tuesday noon, another one tonight at nine pm pacific.
  • Kimberly : What appeals to people about subscription.
  • Paul S : Fits their funding model. Yearly cost. People want a sliding scale. Want more guidance, even though we want to keep it simple. Will get more information tonight and write it up. Everyone is very supportive. Don't know why worm board has trouble with sending emails.
  • Paul S : Someone replied about fees that they're a company and they would be happy to advertise. Ads would be good to push people to alliance. How much revenue could we get, Todd.
  • Todd : From google ads ?
  • Chris : Should only be clickbait.
  • Paul : I'll forward the company.

• JBrowse
  • Paul S : Jordan Ward email wants to have track for knockout consortium variants on jbrowse for Hutter and others insertions for GFPs to use them as balancers.
  • Scott : Don't recall conversation, but we can do that. Need a VCF or GFF or something that says where to get it.
  • Paul S : Can do it with list of alleles
  • Scott : Yeah. If they're already in WB GFF, can easily extract them.

• Nameserver nomenclature.
  • tldr : Use proper names that fit nomenclature for singles, keep using community standard for double and Stavros will see how to much trouble they cause and revisit.
  • Stavros : Thanks to Gary for including Species when curating on the name server. Everyone else, please do that too.
  • Stavros : If something doesn't look like an approved name for Variant or Strain, we're not going to curate it. Strain with lowercase letters is not valid. Alleles with dashes in the wrong place, don't include that.
  • Karen : Any issues with leading zeros in public names ?
  • Stavros : Don't recall that.
  • Chris : Mary Ann said padding zeros in the symbol was against policy.
  • Chris : Can add instructions about that for community curation.
  • Karen : Issue is using the right lab code.
  • Stavros : Not about community, about curators doing that on the nameserver. for example snn-1(S9E)
  • Karen : Was it me ? Was going too fast (or something)
  • Stavros : Yes. An error here and there is not so bad, but it causes problems. Don't have time to be running checks.
  • Chris : I think it was me, didn't realize there was no review process. Thought I'd hand it off to experts and Hinxton. If policy is to omit problematic submissions, we can do that.
  • Stavros : Yes, we decided some time ago not to enter data that doesn't match designations. Can contact authors, or just skip it.
  • Ranjana : Do we need to say elegans ?
  • Stavros : Will default to elegans, so it's not necessary for elegans.
  • Stavros : Another example that should be two separate alleles. jsn13jsn14
  • Chris : We have multivariant alleles that used to be in that format. Have we decided that we don't support that and need a new allele with its own allele designation.
  • Stavros : Yes, that won't go through the pipeline.
  • Chris : What about the several that were already like that.
  • Stavros : I think I removed them.
  • Paul S : Seems like appropriate nomenclature.
  • Stavros : For genotype that's fine, but for allele public name it's not. For genotype that's an understandable mess.
  • Ranjana : User guide says that it's allowed https://wormbase.org/about/userguide/nomenclature#kgjd0fh7ca25m89be4l613i--10 "lin-12(n676n930) n930 is an intragenic revertant of the n676 lin-12 gain of function allele"
  • Paul S : Need to figure out how to handle that.
  • Stavros : I would handle that as two different alleles.
  • Chris : But they're combined into the same locus simultaneously in tandem.
  • Stavros : Create new public identifier
  • Paul S : Doesn't exist by itself, it's too close, could do it with crispr.
  • Chris : Classical variants would show up through mutation screens.
  • Paul S : Crispr is different, keep separate. These are not alleles, they're genotypes that are proxies for real genotypes.
  • Stavros : We don't make a distinction in our model between alleles and variants.
  • Chris : It's a multivariant allele.
  • Stavros : http://www.wormbook.org/chapters/www_nomenclature/caenornomenclature.pdf create identifier and put link to variation (?)
  • Karen : New public name, won't follow standard public name nomenclature.
  • Chris : There isn't an allele n930. Create two separate allele objects, and the Linked_to field to connect them ?
  • Stavros : Linked_to ?Variation XREF Linked_to // for being able to specify paired substitutions which are part of the same allele
  • Raymond : I just added the wormbook chapter link to WB nomenclature userguide widget
  • Chris : Linked_to aren't necessarily unique or exclusive. Can't know unambiguously.
  • Stavros : Good question. We don't distinguish alleles and variants. Things that don't fit schema can go in Remark field.
  • Chris : But it's critical info that won't be seen by users.
  • Stavros : Can phenotype class deal with this.
  • Chris : Not without major remodeling.
  • Paul S : How do mouse people deal with haplotypes.
  • Chris : Don't know. Alliance planning to handle things as Allele objects, and multigenic alleles being handled differently from single. It's hard to unmabiguiusly handle data. Becomes murky.
  • Paul S : Heavy to handle. Should do it like fish people, but since most is simple, we want to be able to map it directly.
  • Karen : Stavros did you suggest making a single object for the double ?
  • Chris : Ideally there would be a separate allele with lab and integer to represent multivariant allele. But the community having done it the other way, do we need a call to the public to tell them about the policy. They can publish what they want, but it the database must fit constraint. Need a supplementary table for the official name.
  • Stavros : Create a WBVar object with a link to both.
  • Paul S : Need to figure out how to munge the official designation.
  • Ranjana : User guide says they must make a genotype.
  • Chris : What does it mean to create a genotype. In WormBase database ? Genotypes are a new thing.
  • Ranjana : I make some for disease. We have an OA to create genotypes. Just saying what the user guide says.
  • Stavros : If we keep making one to one objects, when they see a genotype with two variants back to back, they can deduce they're in the same gene.
  • Chris : It's one allele with two variants, not two alleles in the same gene. We've conflated the meaning of variant and allele in the worm community.
  • Raymond : It's apparent they're of the same gene because it's in the bracket.
  • Chris : Going forward, leave the name blank and add a comment ?
  • Stavros : nameserver might not allow blank name.
  • Chris : Because of format, we cannot create these in the database. Some have a public name that is the WBVar ID. Unclear what we should do in the nameserver.
  • Stavros : WBVar could be public name, but we won't know what the object ID will be.
  • Mark : Nameserver will create the ID.
  • Chris : But can't add the name which is required before automatically create the ID, so they must have come from a different pipeline.
  • Raymond : Strain with multivalue allele https://wormbase.org/species/c_elegans/strain/WBStrain00005527#0317--10
  • Chris : https://wormbase.org/species/c_elegans/variation/WBVar02125351#02-8-10 Something about the variation. Has a Remark to see the original variatn or molucular lesion.
  • Stavros : I'll keep an eye out.
  • Chris : We should follow up later. Don't have clarity on how to submit these to the nameserver.
  • Stavros : Okay. Can try putting it in like that, and we'll see what happens during build.
  • Mark : Hinxton can try to figure this out.
  • Raymond : My example probably comes from CGC.
  • Stavros : Public name will accept any string, even double variant format. Let's keep doing that, I'll keep an eye during the next build about errors and warnings from those.
  • Chris : Thanks.
  • Stavros : Can we try to find this variant with a partial search. Only gets the variation, but not the double. We can enter them, but community won't be able to find on the website.
  • Todd : We can address that.
  • Stavros : n582ad952 = WBVar02125351 Search for n582 returns WBVar00089610
  • Paul S : Genotypes and people are the hardest thing.
  • Karen : Here is another example of how to deal with it: MaryAnn made two different alleles and then an ID for the double: https://wormbase.org/species/c_elegans/variation/WBVar00241524#02-456-10 Been a problem since 2008
  • Paul S : How is crispr handled.
  • Karen : Crispr getting multiple names linked to original insertion.
  • Chris : The main problem is that the community expects to name them this way, but our database wants something else. Going forward at Alliance we can do whatever we want, but it's an issue of reporting and catalogue. Does anyone have standalone names and <something> to start cataloguing. Not a problem that others will have to deal with at Alliance. It's a semantic issue, and policy.
  • Stavros : Double mutant issue might be that I haven't done a lot of double mutants. From these two examples, can see that acedb can handle a string like that.
  • Karen : Quoting Mary Ann, didn't catch it. Made two different variations and made a new one with a name that allowed to have two different concatenations. In Tree Display the public name is the concatenation. Have one object with a double that can connect <something> to.
  • Stavros : Won't be possible to map it.
  • Karen : But can map single.
  • Stavros : Okay, false alarm.
  • Paul S : Need some way to map it, or generate logic for mapping.
  • Karen : Rearrangement should allow larger tracks. But doesn't fit how community thinks about it.
  • Chris : Alliance has distince objects for alleles and variants.
  • Paul : Seems we can handle it, but need to keep monitoring, and Stavros can keep doing that.
  • Stavros : Next time will keep an eye.
  • Stavros : Paulo, we were not able to find Variant objects when doing a public match in search. Sometimes search fails on WBPerson IDs autocomplete.
  • Paulo : Looking into search. Please send example.
  • Stavros : n582ad952 = WBVar02125351 Search for n582 returns WBVar00089610
  • Paul : Project has been around a long time, there's a lot of stuff from the past.
  • Stavros : Also an issue with Is for inserts in Variants, but something about treating them as Transgenes.
  • Karen : Will look for original discussions. There's edge cases. As long as peeople can find them.
  • Paul S : WormBase provides a strain database. When public problems get sorted out (?)
  • Chris : Some people want to create a lot of names, but cherry pick what becomes public.
  • Paul S : It would be private to a lab's portion of the database, and we'd get a big win on nomenclature standardization. Something to think about, not ready to move on it.
  • Karen : Sounds great. Help for natural disasters.
  • Paul S : It has to be private. Push curation into lab records.