wiki:CurationToolMultiGenePhenotypes

Version 29 (modified by mah79, 6 years ago) (diff)

--

Multi-gene phenotype support

Suggestions from Skype calls

Suggested changes from Antonia and Midori

  • Allow editing of conditions after evidence - DONE kmr 2014-10-23
  • Split phenotype table on publication page into single and multi-gene phenotype annotation tables
    • needs fixing anyway as all phenotype annotations appear on all gene pages - DONE kmr 2014-10-23
  • Hyperlink "Genotype name" in the phenotype annotation table - DONE kmr 2014-11-12
  • Under "Actions ..." just say "Add a new phenotype for this genotype" - DONE kmr 2014-06-06
  • On the "Add a genotype" page:
    • add gene names beside the allele names in the allele table
    • add an input field for an optional "name" for the genotype - DONE kmr 2014-06-06 - needs better formatting
    • allow searching by genotype "name" - autocomplete and then fill in the identifiers

New thoughts from testing 2014-10-24 (Val)

  • The Genotype list has instructions "Add a Genotype" I think that for single gene mutant alleles this might be a bit opaque what the user needs to do. I think it might be more obvious if this was called “Genotype and allele list” and the instruction “Add an allele, or a genotype” ...although I am not welded to this and would happily go with any better suggestions (THIS ONE NEED CURATOR DISCUSSION) - MAH 2014-10-28: this 'might' be superseded by ideas from today's call; see below and have a think
  • I think most users won't have a clue what to add for a name. We don't want them typing in the entire plo14-123,cdc2delta,cdc3-345 etc, but we need to make it clearer what to add. So, we need to say (succinctly) something to convey this “This is a optional “short name”, the lab strain name will do, or , another published label to enable to you to find this genotype quickly) “ A ‘formal’ description of the form(plo14-123,cdc2delta,cdc3-345) will be constructed from the individual alleles in the database so you do not need to type that in here" (THIS ONE NEED CURATOR DISCUSSION) - KMR 2014-10-25: It has improved slightly - it doesn't say "click to add genotype name" - MAH 2014-10-28: On today's call we decided changing the wording to "short genotype name (optional)" would do. The documentation we're gonna write any day now will also help.
  • It would be nice if it were clearer when an allele had been added from the genes in your list. Maybe oce you have selected your gens for a specific genotype, once you added the allele the "add allle box" changed in some way (this might not be good if you have 2 allleles of the same gene) (IGNORE THIS ONE FOR NOW)
  • Details should show the formal name (i.e plo1-ts18,nod1delta), and the “user given name” i.e (mystrain2) - KMR 2014-10-28: I had that for a while but it looked like a repeat of the list of alleles, but all on one line. It looked messy for multi-gene genotypes so I removed it. Is there anything we can do to make it look less messy?

  • Choose evidence for annotating x with y -Make it clearer that here we are doing evidence AND condition - KMR 2014-11-12 this should be improved now
    • 1. Suggest Select evidence and add conditions(for x with y)
    • 2. Put a bit of separation between the evidence box and the conditions
    • 3. Previously used conditions do not get added to the list when you click on them - DONE kmr 2014-10-30
  • Duplicating a row, Small bug: When I duplicate a row I can change things, ( change evidence, add conditions) The evidence changed OK, but the new condition I selected did not appear - DONE KMR 2014-11-10
  • At some point (after adding a single gene phenotype) the only button available was “change genotype” I hit this and it took me back to the summary page (unexpectedly, but that was where I wanted to go!). When you curate a “single gene phenotype” like this, the allele name (genotype) does not appear. I think this was because I didn’t name it, but the name should be derived from the allele designation(not from the user) see above. - MAH 2014-10-28: see below for item from today's call relevant to going to the summary page. Other entries may address more of this, but we'll see what needs further discussion or work.
  • Bug, when I tried to create a genotype wth 3 deletion alleles I was not allowed to "finish". I had to "cancel" it instead - KMR 2014-10-25: I can't reproduce this

New thoughts from testing 2014-10-24 (Midori)

  • The gene page needs a route back to phenotype curation (in addition to "change gene"). - MAH 2014-10-28: agreed; see ideas from today's call below
  • The ontology term page needs route back to term search (at the moment the only option is the paper summary page, and the "back" button does nothing). - DONE KMR 2014-11-10
  • The "duplicate" popup has the same header as the "edit" popup ("Edit $annotation for $gene"). Could it say something more helpful, such as "Edit fields to create a new annotation for $gene"? - DONE KMR 2014-11-10
  • The "Add ..." link for the phenotype table needs some sort of cleverness for before any genotypes are added. It would probably be simplest just to have it not appear until after there's at least one genotype available in the session (at least for now). Later we could think about having it link to the genotype-adding interface if there are no genotypes yet. - DONE KMR 2014-11-10
  • In the "standard" phenotype curation interface, the box for adding conditions is subtle and kind of easy to overlook.
    • It could probably be (a) a bit smaller and (b) not rammed up against the evidence selector (i.e. put a little bit of vertical whitespace between the evidence pulldown and the conditions box). - KMR 2014-11-10: I've made it a bit more obvious but we should chat about how it should look
    • Maybe a bit more border, too, although there's a risk that would look naff. - KMR: I haven't tried a border yet
    • It'll also help to mention conditions in the header ("Choose evidence and conditions for annotating ..."). - DONE KMR 2014-11-10
  • It would be nice to have a way to edit a genotype, in case we make a mistake entering details. - MAH 2014-10-28: agreed; see ideas from today's call below
    • Ooh, and maybe also a duplicate-and-edit option analogous to "duplicate annotation"? If I have two genotypes, each with six alleles of which five are the same .. (was Antonia asking for something like this on the last call?)
    • Another thing that might be handy is a list of all the genotypes and their descriptions. It might be too much to put in the genotype box on the paper summary page (some genotypes will be complicated), so maybe a new page or popup, with a "see genotype descriptions" link.
  • Question for curators: should we have a way to delete an unused genotype from a session? We don't want to delete a genotype from the database, but maybe one that gets added in a session but not used could safely be deleted to declutter the interface. - MAH 2014-10-28: see ideas from today's call; see below

Suggested changes from Val

  • have an "edit gene list" on the "Add a genotype" page - call it "add genes" on this page - DONE kmr 2014-06-06
  • "Actions:" should be "Action" on the genotype page - DONE kmr 2014-06-06
  • "Add a new phenotype ..." should be "Add phenotype to genotype" - DONE kmr 2014-06-06
  • Add conditions to evidence page - DONE KMR 2014-11-10
  • on publication page "Add a genotype" -> "Select a genotype" - DONE kmr 2014-06-06
  • on the "Select a genotype" page - MAH 2014-10-28: superseded by ideas from today's call; see below
    • first choose the genes for the genotype - checkboxes - DONE kmr 2014-06-06
    • then "Continue" - DONE kmr 2014-06-06
    • then "choose an existing genotype" from a list, if any
    • otherwise, add alleles ...
  • Add auto suggest existing geneotypes based on the collection of genes selected from the list of available genes (BEFORE alleles are added, you only need to do this if your genetype does not already exist)

To-do items / bugs

  • the same annotation appears on every gene and genotype page - DONE kmr 2014-10-23
  • In the "standard" ontology term search, when I go to select an ontology term, it only "takes" the second time I click it. - DONE, sort-of KMR 2014-11-10
    • Not sure how else to describe it -- click to select a term, and nothing happens; click again and it's selected. (This feels like the sort of bug that's going to be a nightmare to reproduce and track down, but I get it in both Firefox and Safari. Happy to demo if that'd help.) Term selection in popups ("Add", "Edit", etc) works as expected. (mah 2014-10-24)
    • KMR: thanks, it was easy to reproduce but hard to fix. :-) The double click thing is fixed but now sometimes the term detail page doesn't have any detail. Which isn't ideal. I'm trying to fix that now. (2014-11-10)
  • Is the annotation extension box supposed to appear in the "duplicate" and "edit" popups when one is not logged in? It does at the moment ... (mah 2014-10-24) - KMR 2014-11-10: that's definitely a bug
  • I gave my genotypes names, but the "Add" phenotype popup doesn't use them in the genotype pulldown; I get what I presume is the internal ID gibberish instead, and have to guess which is which. (mah 2014-10-24) - FIXED KMR 2014-11-11
  • It quite happily let me add a duplicate genotype. I think we want to disallow that -- cough up a message such as "That genotype already exists [with the name $genotype_name]", don't we? (mah 2014-10-24)
  • Is the definition getting lost for a term suggestion? I can see the name but not the def for suggestions I put in. (mah 2014-10-24) - FIXED KMR 2014-11-10

Ideas from 2014-10-28 call

  • General: any button that goes back to the paper summary page should be labeled (wait for it) "Back to summary". DONE kmr 2014-11-07
    • we'll also think about what other links might be useful, and where hyperlinks or buttons should go
  • Create a separate page for genotype details:
    • Name & description for each genotype used for the paper DONE KMR 2014-11-12
    • "Find a genotype" link - KMR 2014-11-20, this is partly done: there's a find a genotype function but it only searches genotypes from this session.
    • "Add a genotype" link DONE KMR 2014-11-12
    • For any genotypes, add a "duplicate" link that behaves like the duplicate-annotation link (to speed up & simplify adding new ones)
    • For genotypes added in this session, add an "edit" link (spout a warning if you try to edit a genotype that has phenotypes annotated to it)
    • For genotypes added in this session, and NOT used for any annotations, add a "delete this genotype" option
    • KMR 2014-11-10 - there is now a "Genotypes management" page but the formatting is very basic and the bullet points above aren't done
  • "Find a genotype" behavior:
    • Use the "choose genes" interface here -- user selects the combination of genes to search for
    • Search retrieves any genotypes that use all of the selected genes
      • each offers an "add a phenotype to this genotype" link
    • Also offer link to "Add a new genotype for these genes"
      • show the interface with "add allele" just for the selected genes
  • "Add a genotype" behavior:
    • Remove the "choose genes" step from the current interface; go straight to the gene list with "add allele" buttons for each
      • also make sure a genotype can be finished if there isn't an allele for each gene in the list
      • DONE KMR 2014-11-10
  • On the paper summary page, replace the "Genotype list" box with a briefer summary such as "n genotypes" and link to new Genotypes page. Might need further discussion to get appearance and wording nice. KMR 2014-11-12 Currently there is just a link with the text "Genotype management ..." that goes to the genotype page.
  • Gene page:
    • Add list of genotypes involving the gene, with link to genotype page
    • Add back "single gene phenotype" curation type
    • Add message "to annotate a multi-gene phenotype involving this gene, go to the genotype page" (or words to that effect), with a link - DONE KMR 2014-11-12 - let me know if you'd like the wording changed.

Ideas from 2014-11-25 call

  • Search for genotypes by name as well as by genes
  • "Add more genes from PMID:???" is missing from the new evidence selection page

Thought 2014-11-26

I think it was part of the original multi-gene phenotype spec that for genotypes that have both a wild-type copy and a mutant copy of the same gene, either or both can be at the endogenous expression level (if not, it should've been). So we need this change in the interface:

  • Allow a wild type allele at endogenous or "not specified" expression as part of a genotype. The single gene phenotype version should not change -- that should still only offer overexpression and knockdown for wild type. If possible, it would be nice to have a warning if a genotype only has wild type, endogenous expression (and no other alleles) for any one gene.

(prompted by an email exchange among curators)