UHTS_20091113
iPG2P NextGen Sequencing
November 13, 2009, 1pm EDT
Attendees: Matt Vaughn, Tom Brutnell, Justin Borevitz, Karla Gendler, Bob Schmitz, Steve Goff, Steve Welch, Matt Vaughn, Matt Hudson,
Action Items:
Agenda:
Item 1: VizAnalytics Summary
Brutnell gave a summary of the Visual Analytics Mutual Understanding Meeting with the purpose to think out workflows, inputs/outputs. He then presented a workflow that he and Bernice came up with for visualization. Could Justin put together a similar pipeline for SNP visualization?
JB: uses google widgets to view (trait data and their relationships, scatter plots); want to look at linkage and plieotropy; group been focusing on upstream side
TB: after create minimal pipeline for NGS, have plug and play modularity and have the ability to swap in/out different programs. need documentation for iPlant pipeline. real opportunity to set up pipeline that would be go to place
SW: each of the ovals a shopping cart of genes? TB: it is all data but in this particular use case, they are lists of genes
SG: likes idea of making these in workflows; would be good to know on the workflow what requires HPC and what doesn't; also how to coordinate with data integration of other groups such as Gramene or anything Doreen is working on
JB: spatial analysis on the landscape, to integrate phenotype on landscape; climate variables through the seasons from 6 different geographic sites; model fitting and data browsing; visualization should lead back to model generation
SW: switch modeling and literature search boxes so can iterate from 5 genes of interest to modeling box
SW: Eric Lyons suggested we need to come up with generic list types
TB: what's the next step? pass on to other working groups and have them work off of this and to work with Data Integration group
JB: are we going to push for non-reference genomes?
TB: can Justin take lead to come up with a prioritized list of what methods to provide? survey this group and other working groups and find out what is the value of NextGen for other groups?
JB: will take lead on generic species, can you start to assemble that across individuals? Maybe EST assembly;
MV: reviewed status of documents that has put together; under review from Core; what we need next from group is prioritized list; each one of the ideas needs a workflow (need parameters and algorithms); of all the things we could possibly do, what do we want to do?
List:
- non-reference genome and de nove assembly
- new variant detection methods
- etc...
SW: could have white paper, could try in teleconference, or could extract bullets and construct poll for voting
TB: could cut and paste and find email exchanges and could come up with bullet points and use that for discussion in the next meeting
JB: there's some doability to think about; knock it off early or if we put effort in now, will it pay off in the long run
SW: need a process that gives everyone equal input;
TB: we will need to consider input from other groups also
MV: priority list is not binding; we'll have lots of rounds prioritizing
JB: assembling genome of short reads; reduced representation DNA sequencing?
MH: assembly is relatively straight forward for cDNA stuff but don't want to run on a desktop; could do a assembly but average biologist is not going to know what to do with it
JB: assembling short reads, don't really need annotation just want SNP detection;
MH: reagent costs on PacBio machine are way too high
TB: de-nova assembly, thought Pac Bio might be the way to go than with assembly algorithms;
MV: will see reads get longer and algorithms will increase in quality
MH: assembly problem is more biological problem than compute problem (repetitive transposable)
JB: whole genome assembly is one goal; EST assembly is different in that you aren't trying to think everything together but still have separate all ESTs into genes; thinking about reduced representation because get variants
TB: will cut/paste and assemble list of what may need to be prioritized and in next telecon and set up discussion and poll to rank them