New Home

People

VizCookBook

Seelab Home

CS Home

UTK Home

 






SeeGraph Homepage

by Joshua New



           


I) What is it?
SeeGraph is a package for the interactive visualization and analysis of weighted, undirected graphs and is capable of handling graphs with thousands or even tens of thousands of vertices. It renders graphs using the common node-link scheme with quadrics and splat halos for enhanced depth perception as well as in compressed adjacency matrix view. It consists of approximately 15K SLOC implemented with efficient C/C++ and OpenGL. It can be built and run in either Linux or Windows and the eventual open source version can be built in Linux using make or in Windows with the workspace (*.dsw) or solution (*.sln) files.


II) What can it do?
Algorithms contained include several common graph-theoretic algorithms, 2D and 3D versions of Kamada-Kawai and Fruchterman-Reingold and Energy-barrier graph layout algorithms, block-tridiagonalization for permuted ordering of blocks along the adjacency matrix diagonal, interactive compound boolean range querying, and dynamic neural network filtering. The package has been designed to be extensible with built-in capabilities for loading custom layouts, calling other packages (such as R or SAS) at runtime, and other import/export capabilities. For more details, see the following information:
  a) New, Joshua R. and Kendall, Wes and Huang, Jian and Chesler, Elissa (2008). "Dynamic Visualization of Co-expression in Systems Genetics Data." In IEEE Transactions on Visualization and Computer Graphics, Vol. 14, No. 5. [pdf preprint]
  b) SeeGraph Intro [18MB PPT] - short introduction of capabilities with lots of pictures


III) What does it look like?
SeeGraph has applied general graph theoretic analysis and interactive visual analytic capbilities to several domains including:
  a) Conference videos:
    i) TVCG - supplemental material video highlighting the usefulness of various capabilities.
    ii) ISBI (TSCC codec required) - Institute for Systems Biology demo video of loading SeeGraph with the new interpreter goose.
  b) SciDAC posters - VERY large pictures for demo on ORNL's 30'x8' powerwall EVEREST:
    i) SciDAC - Combined poster of genetics and proteomics applications with explanatory system diagram.
    ii) Genetics - example of genetics workflow with sample of results including paraclique visualization.
    iii) Proteomics - SeeGraph applied to proteomics pathway analysis.
  c) SciDAC research highlights - small picture and short summary of various applications involving SeeGraph:
    i) Dynamic Visualization in Genomic and Proteomic Research of Yeast for Bioenergy - yeast research for renewable energy.
    ii) Dynamic Visualization of Coexpression in Systems Genetics Data - microarray analysis of cerebellum development in mice.
    iii) Pairwise Axis Ranking for Parallel Coordinates of Large Multivariate Data - optimization problem of ordering axes in parallel coordinates.
  d) Picasa Album - an unsorted collage of SeeGraph pics and related technical diagrams.


IV) Can I use it?
An old version which has many bugs is all that can be made available at the moment. With that warning, you can get the following:
  a) Download - for Windows XP, Debian Linux, 64-bit SUSE Linux.
  b) Input Files - description of file format for various functionalities SeeGraph suports with links to example files if you just want to try it quickly.
  c) Usage - how to interact with the program, including a handy cheat sheet for key commands.


V) What has it been used for?
SeeGraph has applied general graph theoretic analysis and interactive visual analytic capabilities to several domains including:
  a) Genetic analysis - introduced and described in "Dynamic Visualization of Co-expression in Systems Genetics Data" published in IEEE Transactions on Visualization and Computer Graphics [pdf preprint]
  b) Material science - interactive visual analysis of over 1 million particles
  c) Image processing - animated gifs of results from SeeGraph(v0.40) with data and description of capabilities















        

Suggestions or Collaboration?

Visitors since Sept. 4, 2005:
SeeLab, Computer Science Department
The University of Tennessee
Knoxville, TN 37966

Phone: (865) 974-5400
Fax: (865) 974-4404