Notes from iPlant conference

May 24-26, Las Vegas, NV

Conversations / meetings

What conversations did we have about treeviz at the conference? Interesting use cases, desired links to other resources, feature requests for viz?

Karen's notes about linking data to trees

Many of the comments were of the form 'map data onto tree and then link to analysis / model development', which indicates that treeviz will play a role but cannot fully satisfy this research need. Other WGs required.

Eric Lyons - links to genomic resources. Given species X, which closely related species have genome data? What quality is genome data (finished, partially annotated, assembled, depth, etc). Has resource with APIs at Berkeley. Would be interested in embedding viz tool.
Thomas Brutnell - C3/C4 transitions
Angus Murphy - links to protein 3D structure. Structural change events on tree; know relationship between molecular sequence and structure changes; help to develop models for structural change.
Jeff White - visualizing pedigrees, networks instead of trees; have detailed pedigree information about crop species going back to 1800s.
Andrea Schwarzbach - biochemical pathways; which species use which pathways, often more than one, or use part of a pathway.
Richard Condit - uncertainty, both in terms of annotations on trees (e.g. bootstrap values) but also alternate topologies, different methods needed at different scales. Two trees vs 10 trees vs 10000 trees.
James Macklin - API for user-driven annotation, with strong links to ontolgies without requiring this knowledge from users
Mark Schildauer (and others) - search for list of species (community) and show subtree or prune tree to display only those communities

Other comments

A few people were concerned about a focus on The One Big Tree. Need to communicate that building The Plant Tree of Life is only one part of this project, and that the tools and cyberinfrastructure will be available for other data sets (from databases or user-submitted).

Bernice's notes from the conference

1.  One prominent computer scientist was concerned about how robust and sustainable our infrastructure was, and how it is different from other cyber infrastructures.  If we are similar, are we leveraging others?
 
2.  Some people commented that we are focused in our infrastructure on size and throughtput.  Are we offereing any special functionality for plant biologists?
 
3. There is a lack of clarity on our modular CI apporach, where the community can build on top of existing infrastructure.  How to integrate with existing work?  What is required to integrate with projects such as iPToL tree visualization?

4. Lots of support from some smart people (e.g., Malcom Bennet) on the value of linking multiple data sources and components.

5.Tree of Live Visualization-  some people thought it was not useful to most plant biologists since they only care about one clade.  There was interest in combining data from multiple experiments onto the Tree of Life Visualization.  Could this be done as a community using social software approaches?

6. There was confusion about what exactly we were creating and how they can contribute.  Some people thought we were just creating tools (e.g., Tree of Life Viz).  Some people thought we were just providing servers and porting code, etc. 

Next steps for viz

Karen:  Now that we have a scalable viz tool, the next big feature is linking the viz app to metadata (either through file uploads or online sources). How should we go about developing the API? This will require our group to integrate with other working groups such as iPTOL Data Integration, Core Software. Are there specific people that should be involved? Bill Piel and Val Tannen from iPTOL, definitely. Who from Core Software?

Bernice:  I think it would be opportune timing to create a web service that would support mapping of metadata from one visualization or analysis onto another.   Of course, our first test case would be to support mapping metadata onto the Tree of Life Visualization, and mapping selections from the tree onto other visualizations.  But, the goal would be to build a general service that could be used throughout iPlant.  Adam and I are beginning discussions about such a tool.