TR_01MAR10

Call-in/WebEx Information

  • TR_WebEx - this meeting will include visuals, so WebEx Video will be utilized

Action items

  • AI: Former user (Deleted) will present rough UI for tracking results/hits [incomplete]
  • AI: Former user (Deleted) to mockup an expert tree prune/selection w/ auto-complete to find taxa. Common taxa box should be incorporated. APWEB has an example of how to show users where their species/group of interested is placed on the tree. [incomplete]
  • AI: Include this graphic "progress" shown in steps/blocks for "Search" & "Single BLAST" [done]
  • AI: Change "Gene Term Search" -> "Search" [done]
  • AI: Should remove the curly bracketed format/syntax explanation - not needed. [done]
  • AI: Gene Context should be changed to "Gene Focus" [done]
  • AI: in Use Case #2, change "reconcile" to "submit" [done]

Agenda

  1. Brief review (10 min.)
  2. Prototype Progress Update (10 min.)
  3. Finish up initial review of UI Mock-Flow for Phase 1 Prototype (remainder)
    • Feb. 15, left off discussing Use Case #3
  4. Admin Note:
    • Next Meeting: March 15

Open Questions

  1. Regarding the creation of an iPlant Gene Catalog, what should be included beyond 1KP or 1KP Pilot Data?
  2. Have we decided on 'gene cluster' over 'gene family'?
    • [Answer] The term 'gene family' has some assumptions held by users in the targeted group that these genes have the same function. There are no assumptions around the term 'gene cluster' or 'cluster' so usage of this term would be more clear. (source: discussion between Todd & Cecile)
  3. Regarding the details included in results views:
    • Do have a complete list of what needs to be included at family/cluster level re fields/metadata?
      • [Answer] Size, # of species, annotation, domain, GO term
    • Do have a complete list of what needs to be include at the individual gene level re fields/metadata?
      • [Answer] Species, Definition Line, Aliases
  4. It was suggested that if the entire subfamily was not select (so particular members were checked in the checkbox interface, but not all) we might want to use some visual representation of this.
    • How do we represent subfamilies or families not checked by the user?

Notes

Attendees: Former user (Deleted), Sheldon McKay, weix, ane (Unlicensed), tjv, Former user (Deleted), NicoleH

  • Review
    • Two Action Items assigned to Andrew L. remain incomplete.
    • An overview of the discussion between Adam Kubach, Andrew Lenards, Andy Muir is available for review.
      • the focus of discussion centered around tree visualization requirements current being discussed in our working group (TR)
    • Sheldon attended the Tree Visualization Workshop (in Austin Feb. 25 & 26, 2010): the point was made reinforcing the cross-cutting nature of the Tree Visualization Working Group. Some functionality will come from their broader plans, but the main discussion regarding Tree Reconciliation needs was 'fat tree' representations & how they will scale (still needs to be determine if 'fat tree' will remain useful for very large reconciliations).
  • Prototype Update
    • Sharon is helping Sheldon with pre-computation backed (TreeBeST/PrIMETV).
    • Andrew L. will begin UI development soon.
  • Mock-Flow Review (Focus on: Use Case #3)
    • We have a user interested in Arabidopsis & Poplar. They curious about the events (gene duplications, in this case) leading to those two groups in NCBI taxonomy tree.
      • Clarification: the user got to this point by performing a search - it is not currently clear in the mockup where they came from [consider-change]
    • [User Action] the user clicks the branch leading to the most recent common ancestor (MRCA) for the two groups and then is taken to a results view w/ tabulated information.
      • [open question]: what should be between the description & #dup/total columns?
    • Clarification: one column is "#dup/total" and values are shown: 8/60
      • what does the ``total`` represent? In this case, the 60 is the total number of events on the branch. Another way to put it (per Todd), All duplications from all gene clusters on that branch and only that branch. Thus, all entries for ``#dup`` should add up to that number. The ``#dup`` represents specific number of that cluster as opposed to that GO term.
    • Todd suggested considering a way to filter these results based on description. Or other attributes to limit the results. There will be an enormous number of results of all clusters represented on branches, how do we let the user focus on what they interested in/need? Filter out over-represented clusters? Filter by function? Perhaps rank by the # of events in that cluster relative to the branch?
    • Consider this:
      • Denominators/Statistics:
        1. all events on branch for that cluster
        2. number of events in that cluster across all branches that the cluster appears
        • Then a user might look at top & lowest ranked families/clusters
        • These two statistics becomes a way for families/clusters to stand out
          • So a user clicks on a branch and looks at these statistics for a way to narrow what they're after
      • Also, a user might search for ``globin`` (shown in mockup) and select Gene Family/Cluster #79 (shown) and then want to see all branches that this cluster (#79) appears on the "big tree" (or NCBI taxonomy tree for interim).
      • For the prototype, focus on events representing gene duplications
    • Andrew L. asked about how adding event types might change what is shown (say losses are included, or horizontal gene transfer).
      • There is no way to include these things all at once right now because different models are used to infer each event type and a mixed-model approach has not be developemd
        • A subset of species w/ complete genomes that you can calculate losses for could be used - but won't work for those without complete genome.
        • Are users interested in losses that are based on lack of data or missing data? Yes.
        • AI: need a list of issues to deal w/ in later versions (create this in wiki)
          • AI: add to this list: deal w/ missing genes and losses
    • Showing User Progress through Steps
      • Andrew L. added rough abstraction in the mockups
      • an abstract representation of some set of arrows to specify walking through workflow from the beginning to the end (and where they are in the process)
        • Would like to be able to specify the steps in some discrete way with descriptions so that mouseovers can be viewed by user (select branch -> select cluster of interest -> view reconciliation).
          • Ideally defined in steps from the entry point in a high level way.
    • 3 panel view
      • add some details to describe what the black/blue zoom position box is.