TR_15FEB10
Note: No meeting scheduled for Feb. 22, 2010
Call-in/WebEx Information
- TR_WebEx - this meeting will include visuals so WebEx Video will be utilized
Action items
- AI: Andy will present rough UI for tracking results/hits
- AI: Visualization needs of TR working group need to be communicated to Tree Viz working group [done, Tues., Feb. 9]
- AI: Make available information about Todd's suggestion of using Linked Hyperbolic Trees as a way to visualize reconciliations [done, Tues., Feb. 9]
From the Tree Visualization Working Group
Sheldon and Andy sat in on the Feb. 9 TV working group to present the needs of TR. Karen Cranston suggested that both Sheldon and Andy review TV Use Case #2 regarding the comparison of trees.
Agenda
- Brief review of NESCent visit (5 min.)
- Discuss In-scope Use Cases (#1, #2, #3, #12)
- Review rough UIs for Use Cases (#1, #2, #3, #12)
Open Questions
Notes
Attendees: Natalie Henriques, Adam Kubach, Sheldon McKay, Todd Vision, Jerry Lu, Jim Leebens-Mack, Nicole Hopkins, Andrew Lenards
- Meeting started with an impromptu discussion regarding visualization primarily between Sheldon, Adam, & Andy
- The idea of a 3-panel view consisting of a gene tree viewer, species tree viewer, and reconciled tree viewer was introduced. This was a development of the Week of Feb. 9 NESCent visit.
- Specifically, the ability support zooming and panning for a static image (which could be produced for reconciled trees by PRIMETV) was introduced. Several Flash viewers support this interface. Two viewers, in particular, were discussed by Todd & Andrew. Adam mentioned that there are many such viewers.
- Discussed halted when all members were gathered.
- Discussion of prototype and UI mockups begins
- 2 entry points (Single BLAST & Search [formerly referred to a 'Gene Term Search'])
- Single BLAST
- it support the selection of blastp or blastx
- Why? The iPlant Gene Catalog will be storing protein sequences, blastp is used for 'protein to protein' searching and blastx is used for 'nucleotide translated to protein' searching (Andy answering this question from Nicole)
- The iPlant Gene Catalog is the 1KP (aka 1KPT = 1KP Transcriptome data) plus some additional inclusion of data (open question: what should be included beyond 1KP or 1KP Pilot Data?). The pipeline for producing 1KP Transcriptome data will start with "pilot data", Todd said that the 1KP Pilot Data will include data for about 40 species. We should have access to the pipeline that will be producing the 1KP Transcriptome data.
- BLAST mockup includes the initial parameters that users will want to change (scoring matrix, expect values), but is not complete in the form shown.
- it support the selection of blastp or blastx
- Search
- {This is/was referred to in notes and original mockups as "Gene Term Search", but discussion led to suggestion that it be truncated to simply "Search"} AI: Change "Gene Term Search" -> "Search"
- Some value is entered into search box -> "all" from drop-down allows for search against all links to database or "terms" mentioned in the drop-down to determine what the user is looking for without them having to explicitly specify which term/identifier they seeking. The drop-down is then used for optional narrowing by expert/experienced individuals that know exactly what they're looking for.
- The 'Info/Help' hyperlink shown is meant to provide users not family with a particular term/identifier context/knowledge about the formats expected (AT#####, GI######, OS####, etc).
- AI: Should remove the curly bracketed format/syntax explanation - not needed.
- Missing drop-down values: annotations & definition line (Jim), InterPro/PFam domains (Todd)
- Clarification: the term "annotation" related to records and entries is referring to all metadata present. Not necessarily a specific "annotation" entity (like someone from a Computer Science perspective might think). The Definition Line is a specific example of an annotation (or metatdata) for a GenBank record.
- Discussion re Reconciled View
- Looking at showing a reconciled view in a viewer like MorphBank's FSI Viewer.
- 2 options for for viewing reconciliations
- fat tree
- "decorated" gene tree -> a gene tree w/ duplication events attached/illustrated (but annotations/etc)
- Single BLAST
- 2 entry points (Single BLAST & Search [formerly referred to a 'Gene Term Search'])
- Workflow ("Gene Term" Search)
- In the results, the "annotation" field mention is info about the genes in a gene family.
- Good example is Berkeley Phylogenomics
- In PlantTribes - had query of all definition lines from all genes & rules for chosing most common terms.
- AI: Think about rule for annotations included (unassigned)
- User could look at more results linked to from the "in-widget", tab included results
- "more results" takes to a Phytome-inspired view, info about gene family (open question: do have a complete list of what needs to be included at family/cluster level re fields/metadata?)
- under the gene family/cluster level, info about each gene (open question: do have a complete list of what need to be include at the gene level re fields/metadata?)
- Each gene is hyperlinked to the external sources pertinent to that gene/organism.
- selection of genes gives "Gene Context" in next view (3-panel-view), which equates to highlighted node (or markers) in the reconciled tree.
- AI: Gene Context should be changed to "Gene Focus"
- [REVISIT] Focus on taxa - a user just wants to be able to focus one or two species out of the whole species tree. So a desire to narrow the number of species shown in gene tree.
- Chose taxa to include in the gene family views
- This creates a 'pruned phylogeny' - {comments} this gets complex very quick, we'd have to have some tutorials available for how to navigate through or narrow to the number of species the user cares about.
- "Make it intuitive" - How would our targeted users expect to 'input', or specify, this pruned phylogeny? (<-open question) ... Todd & Andy discussed an expert, taxonomic selection approach (like a collapsible tree list) during the NESCent visit. Such a hierarchical "clicker" would be great for selecting taxa but becomes difficult when selecting branches.
- Targeted users not overly familiar/comfortable w/ phylogenies.
- Searching on higher taxa or some way to get to those species easily
- Maybe include "most commonly" used taxa in a box - this might appeal to users from the big model organism projects. Further point: should show where model organisms fall on the big tree (or NCBI taxonomy tree).
- AI: Andy 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.
- Chose taxa to include in the gene family views
- In the results, the "annotation" field mention is info about the genes in a gene family.
- Workflow (Single BLAST)
- Does "Gene Focus" solve the need to track hits (include & selection)? No, it's solving a different issue
- Tracking hits is about allowing visualization of genes & what clades they are falling in on the gene tree.
- Use of word "family" is controversial -> what is the best way to describe this concept for our targeted user group?
- family/clade/subfamily?
- cluster/subcluster?
- Jim likes idea of chosing genes and highlighting on gene tree where the genes reside.
- Gene X (used in mockups) resides on the gene tree, need about to show all species where this resides (in species tree)
- need a way to chose subcluster (subfamily), propose a "next page" that shows species tree for selection
- Missing a way to narrow this selection
- need a way to chose subcluster (subfamily), propose a "next page" that shows species tree for selection
- Gene X (used in mockups) resides on the gene tree, need about to show all species where this resides (in species tree)
- Jim saw a potential issue: "Would a user know why they're selecting genes?"
- Some text description to define why is needed perhaps?
- Todd suggested that "progress" be shown in a flow using steps/blocks, this would indicate the steps being taken and how close the user is to the end
- so a graphic indicating the workflow would be used (a more cartoonish, coarse-grain view acknowledging a 'wizard' set of actions)
- AI: Include this graphic "progress" shown in steps/blocks
- Jim suggested: "A cartoony additive representation of a gene could be used and built up as the user includes gene"
- Does "Gene Focus" solve the need to track hits (include & selection)? No, it's solving a different issue
- Use Case #2
- What's the entry point for this? (Andy)
- The Search is a natural entry point, but either of the 2 identified entry points (Search & Single BLAST) should be valid.
- communicate verification for submission
- AI: in Use Case #2, change "reconcile" to "submit"
- The Search is a natural entry point, but either of the 2 identified entry points (Search & Single BLAST) should be valid.
- What's the entry point for this? (Andy)
- Left off discussion at Use Case #3 mockups
- Jim wants to prepare a powerpoint w/ some ideas/suggestions for the user interface.