Version 39 (modified by kono, 16 years ago)

--

Satellite meeting for Data Visualization

Topics

Visualization for large datasets including:

  • Protein-Protein Interaction
  • Pathways
  • Genome
  • other

This includes development of easy-to-use backend data services such as

  • Interaction network data service
  • Annotation service

Chairperson

Participants

  • Bruno Aranda (baranda@…)
  • Keichiro Ono (kono@…)
  • Soichi Ogishima (ogishima@…)
  • Kozo Nishida (kozo-ni@…)
  • Yasukazu "yaskaz" Nakamura (yaskaz at ddbj.org)
  • Shinobu Okamoto ( so at kazusa.or.jp -- GenoDive)
  • Kunihiro Nishimura (kuni at cyber.t.u-tokyo.ac.jp -- GenoDive)
  • Isaac Ho ( iyho@… -- JGI / UCSF )
  • Yunsun Nam (namy002@…)

Targets

  • Last yeat in the hackathon, we have developed a common query web service API for interaction databases called PSICQUIC. Some of the service providers have published their service and now we should think about how to use the service in client and make them easy-to-use for all service users.
  • The goal of this group is the following:
    • Make backend service more easy-to-use client software
      • Cleaner API
      • REST support
    • Create intuitive UI for client software
    • Design more sophisticated visualization

Date

  • Most of the time during the hackathon. Feel free to join anytime :-)

Room

  • Chura Hall (3F)

Presentations

Notes

  • Data Provider
    • Interactions
    • Pathways
      • KEGG
      • Reactome
      • BRENDA
    • Annotations(service)
      • Biomart
      • Entrez Gene
  • Large-Scale Data Visualization
    • Backend
      • Import data - PSICQUIC, DAS-MI
      • Edit data in database interactively - how?
  • Interactivity

Results

  • Generic ID mapping for GenMAPP Data Import plugin ( by Isaac Ho / UCSF )

GenMAPP-CS is a next-generation implementation of GenMAPP, a metabolic pathway visualization tool, as a set of plugins for Cytoscape. This week I worked on creating a way to map the ids of any network based on GenMAPP-supported data libraries ( e.g, KEGG, Affy, Ensembl ) to the ids of any expression data based on these libraries. The solution involves converting all ids to Ensembl IDs upon import by calling a webservice on GenMAPP's servers which query their MySQL dbs. Kei Ono and I discussed issues such as the best place in the plugin code to place the responsibility for id conversion. The current status is about 60% complete on this task.

  • PSICQUIC: Bruno implemented a REST version of the PSICQUIC service. It returns the standard PSI-MITAB format.

TODOs

Rendering

  • Build a flexible toolkit for biological data visualization. Some of the works above are done by Processing and it is a good candidate for building such toolkit.
    • OpenGL java library (JOGL) should be converted to OSGi for Cytoscape + Processing for large-scale visualization to utilize hardware-acceralation.
    • Integrate Processing rendering engine to Cytoscape for more flexible visualization, including animation, 3D, etc.
    • Genome-Browser also can take advantage of this type of toolkit.
      • What are the common components we need?
  • Visualize time-series microarray data using Cytoscape
    • Animation (using dynamic layout)
  • More biologically meaningful graph layout
    • List up several methods of graph layout

Backend Data Services

  • Make BioMart Cytoscape plugin into SOAP/REST hybrid for nicer GUI and performance. Still Biomart SOAP service has problems if developer uses Java code generator (WSDL2Java/CXF)
  • Use DAS and other services for streaming data for visualization (just like genoDive does).
  • PSICQUIC service is now available as REST and SOAP. We need to promote this standard to the interaction data service providers.

Attachments