All-hands breakout session notes

1
All-hands breakout session notes

• March 2002
• Gatlinburg TN

2
Summary

• What is our product
• Web based service geared towards biologists
• Workflow based interface
• Not providing a toolkit
• Long-term / community support a real issue to be
  resolved in the future
• Targeting small number of users in DOE collab
  institutions
• After 1 year (prior to review)
• Focused on single workflow for Matts scenario
• Workflow interface hosted at NCSU
• After 3 years
• More general workflow capabilities

3
Summary

• Demo details
• Chip data (download initiated by Matt)
• ClusFavor gt list of genes
• Wrapped (by NCSA?)
• User selection of genes gt gene accession s
• NCBI keyword search gt gene sequences
• Eventually need user input here
• Blast search against HTGS gt alignments
• User selection of sequences gt 2000bp seq
• Transfac / MatInspector gt ???? (ask matt)
• User input
• What do we need to do with final output ??
• Alternative scenarios working through this
  workflow others

4
Summary

• Schedule for Demo Sept 9 (all hands meeting)
• Email data / reports through mail list to be
  archived
• Know if we will be ready by all-hands
• Plenty of time for Matt to use it before DOE
  review
• NCSU
• WSDL description of wrappers and composite w/f
• June 15
• Workflow interface
• Aug 15
• Requires obtaining input / output / flow for all
  wrappers and composite w/f (in XML format) no
  later than April 15
• Will provide links to existing demos in near
  future

5
Summary

• SDSC
• Workflow scenario
• August 15
• Will use XWrap generated wrappers to access all
  sources if wrappers delivered by June 1
• Interface
• July 15
• Specific interface to obtain required user inputs
  and adjustments in a reasonable fasion
• Need intensive interaction with Matt
• Additional queries should highlight aspects of
  semantic integration

6
Summary

• GT
• Wrappers
• June 1
• 5 or 6 wrappers needed depending on Matts req
  for output
• Expected to be hand coded using scripting
  language (which is then fed to wrapper generation
  code)
• Matt
• Get example input data sets to us ASAP
• Describe how he needs the output of the workflow
• Intensive meeting with SDSC ( others?)
• Bridge funding? Travel funding?
• Sooner the meeting the better
• 2-3 days 3 hours a day (more or less)

7
Summary

• Implementation details
• Wrappers deployed on server side
• Return results in XML
• Object code linked in to integration component
• Currently hidden by SDSC interface
• Agent interfaces developed by NCSU
• WSDL / SOAP
• Used to connect agents (wrappers) located on
  distributed sites

8
Summary

• Repositories
• Wrapper repository at GT
• Currently in place
• Not password protected for now (may change in the
  future)
• Code locations
• Wrapper generator at GT
• Wrappers transferred to SDSC (via repository)
• Query engine at SDSC
• Interfaces
• SDSC will host poor mans interface to workflow
  scenario
• NCSU will host workflow based interface and
  connect to SDSC service using agent technology
• Password protected ( SDM / 1SDM)