TR_NESCent_Visit_09FEB10_Proto_UseCases
Please note: more information regarding the numbered use cases is here
Use Case #2
Actions/Steps:
- user selects a gene family (entry point)
- selects subsets of genes within that family
- display reconciled tree (use 3-panel view or similar approach)
Note: the duplication events may have annotations "on them", the most valuable being those that indicate 'whole genome duplication.'
Use Case #3
Background: whiteboard
- User: Evolutionary Biologist
- Interested in the most recent common ancestor (MRCA) of arabidopsis & populus
Actions/Steps:
- The user views the NCBI tree (entry point)
- Select the branch leading to arabidopsis & populus
- The user is shown a page w/ "events" related to the selected branch
- This should show the user information like: family ID, description, GO function, # of duplications / total genes
- Question: What fields should be shown in this list/results-page?
- The user can select a family and view a reconciled view (the 3-panel view approach or similar)
Use Case #12
Actions/Steps:
- The user selects "interested" gene families
- The user is shown the available tree (re prototype NBCI tree) w/ additional information on branches
- The tree shows condensed representation of statistics related to a branch
- statistics per branch: gene family copy #, total gene #,
- a suggestion of condensed representation below
- a pop-up also shows details of the duplications present, one only present here, and potential losses
Use Case #4 and it's relation to Use Case #3
Todd suggested that Use Cases #3 & #4 have a relationship in that they're like a faceted search [whiteboard]. The point of #4 is that narrows the search to a pathway related to a GO category.
So the entry point for #4 is really a GO Category.
A "clickable" tree viewer would be required to handle the user interaction that was discussed. Attached information to the branches of the tree would be available and the user could be presented w/ a popup/bubble containing further details about this information. From this interface, a user could select to see a "reconciled" view (which would be the 3-panel view or similar approach).
A second approach, still clickable, could be done w/ a tabbed view that toggled between the species tree (w/ attached info) and the 'details' corresponding to the attachments in the other tab.
Given the scope of the prototype, such a "clickable" tree viewer is not likely. An approach of 'labeling' the tree so the user would indicate/select "branches of interest" in an interface below the tree.
It would be nice if the user could prune the available tree (NCBI taxonomic tree or "BIG" tree) and control the level of detail shown:
[another angle of the above image]
A user then would be able to include or exclude various branches and control the species tree shown. Now, the details that are "attached" to the tree will already be present in the data-source and, thus, will not require re-computing or re-execution of the tools or analysis (or that's at least the idea at this point).
The target user group for this prototype (e.g. Discovery Environment perspective) will not be overly familiar with phylogenies and phylogenetics. They are looking for the phylogeny to help explain the protein or pathway they are working on. So it may be helpful to provide a sort of expert interface for selecting a subset of a tree via a hierarchical interface (this should be familiar to the targeted user group).
Some ideas on how this might be done were discussed. An example of the hierarchic interface from Phytome was shown:
