SC_20110614

iPG2P Steering Committee

June 14, 2011

Participants

Agenda

• General Business (5 min.)
• Modeling framework (Steve W, 20 min.)
• Cornell trip (Matt V, 20 min.)
• Demo of TiTo-based tool integration (Matt V, 5 min )

Notes

General Business

• Tom is moving to Danforth Center in January. 
  • Jim Carrington is the new director.
  • Danforth wants to have a focus on computational and system biology. Want more grass focus too.
  • This presents an opportunity to increase iPlant’s interaction with people at the Danforth Center. 
  • Todd Mockler at Danforth is in contact with Ann Stapleton and Matt V.
• Meetings will be monthly during summer. 

Modeling Framework (Steve Welch)

• Steve's slides
• CM: What does the G2P extender tool do?
• SW: Give it a model. Give it a .ini file. It takes the data and takes multiple subsamples of it. For each subsample, fits the model based on the constraints supplied. For example, for 2000 resamples, it keeps track of which subset of the data gave a particular result. R script asks if the parameter set that gave a particular result is stat diff from others. Sub sampling must be stratified with respec to the role of the variable in the model.

Matt's Cornell Visit

• At end of April, he visited Cornell.
• The visit kicked off a number of interesting initiatives.
• He met with 
  • Tom’s group plus C4 people who were visiting
  • Ed’s group Qi
  • Chris Myers
  • Lucas Muller
• Meeting with Tom
  • Tom proposed the Gates C4 community take advantage of some of iPlant ‘s infrastructure; possibly a persistent, powerful Atmosphere VM.
  • As tools and algorithms developed in such a VM, it should be relatively trivial to push them to DE.
  • A more immediate deliverable is an RNA seq pipeline developed by Lin in Tom’s group.
  • Objective is to deploy it in DE at same time the methods paper is published.
  • Looks promising to have it available by ASPB as an alternative to Cufflinks.
  • Tom will make sure Matt has most recent code.
  • Lin wants to submit the paper this week.
  • Cufflinks can’t deal with strand-specific data sets. Lin’s scripts can.
  • Matt wants to be able to highlight how iPlant enables publication of code coupled with an environment in which it can be used (in contrast to just releasing the code).
    • Peak Ranger from one of Lincoln’s group is in DE.
  • TB: when code is contributed, is it “required” to be documented?
  • MV: We encourage people to document it but can't force them.
• Meeting with Qi (Buckler group and Director of Cornell Bioinformatics)
  • Wants to take advantage of iPlant's HPC Foundational API.
  • It was in feature freeze, but thawed (added some code), so are a bit behind.
• Meeting with Peter Bradbury and Chi Wazu (?) from Ed’s group
  • They want to push the command line version of Tassel into DE.
  • Once Foundational API and DE work well together, we’ll build an interface for it in DE.
  • Genome selection, genotyping by sequencing will also be in later this summer.
• Ed wants to teach some classes on genotyping by sequencing. 
  • Matt is optimistic this can be done in DE this fall.
• Meeting with Chris Myers
  • Sloppy Cell – explored if some of it’s workflows could be in DE.
  • Dependency issue needs to be resolved.
• Meeting with Lucas Muller
  • Has been given access to large memory machine with 1 terabyte for tomato and solanace assemblies.
• TB: There is enthusiasm for adopting the iPlant infrastructure. In August Xing Wang will visit. Will things be ready?
  • MV: VM’s can be set up any time -- after ASPB.

Demo of TiTo tool integration tool

• Last week was the TiTo debute of how to build tools in iPlant environment and user acceptance testing.
• Matt live demo’d how to integrate a tool using TiTo.
• Currently need iPlant support staff to install the app on the cluster. That won’t be the case in future. App can be in your environment on the DE.
• About a dozen tools of varying complexity got integrated last week.

Miscellaneous

• Next Meeting July 19th (a week later due to BSA)

Action Items

• Tom will make sure Matt has the latest version of Lin's code.
• Martha will reschedule the next meeting for July 19.