Documentation

Contents



How to get started

The 3D gene-condition-time microarray date clustering tool EDISA is available in two different versions. If you want to use EDISA right away, you can use our webservice, which does not require the installation of the tool on your computer. If you prefer to install EDISA locally, you can find the latest version of the tool at our download section.

Installation

After downloading the latest version of EDISA, unzip the file to a directory called "TwoStepClustering-Beta".

Installation on Windows without Matlab
  • Download MCRInstaller.exe from the download section
  • Double-click MCRInstaller.exe to install Matlab Compiler Runtime
  • Double-click EDISA.bat to start EDISA


Installation on Linux with Matlab and statistics toolbox
  • Start Matlab™
  • Add the installation directory of EDISA to the Matlab paths by selecting the option File -> Set Path in the menu bar
  • Change to the installation directory of EDISA
  • Type ./guiEDISA to start EDISA.


How to use the tool

  • Start EDISA
  • Load/select a dataset
  • Chose a module type and specify the parameters τG, τC and "number of iterations" (see section below)
  • Click on "Run EDISA"
  • Plot the results by clicking on "Plot module" or "Slideshow"


Parameters

The user can specify the three parameters τG, τC and the number of iterations and additionally choose an appropriate gene module type to be searched for by EDISA.

Correlation thresholds for genes (τG) and conditions (τC)

τG specifies how well each gene has to be aligned with the average trajectory of the module and τC specifies how well each condition has to be aligned with the average trajectory of the module. Low values of τG and τC will create modules with few highly correlated gene expression profiles. Increasing the values of τG and τC will result in modules containing an increasing number of genes that display a reduced correlation.

Number of iterations

The number of iterations are samples, which are iteratively drawn from the dataset. Values between 1.000 and 10.000 are recommended, depending on the size of the set of differentially expressed genes.

Module Type

Three different types of gene modules can be found by EDISA, each of which are derived from a particular biological intuition. Single response modules associate genes with one condition, uncovering very specific response mechanisms that may help biologists to find marker genes for certain stresses. Coherent modules, on the other hand, reveal co-expression under multiple conditions and display a more general response. Independent response modules allow for a more complex type of modular co-expression, i.e., they hint at the existence of stress response specific for every condition alongside with a common transcriptional control.



Adrian Schröder
http://www.ra.cs.uni-tuebingen.de/software/EDISA/intro.htm
© 2008 University of Tübingen, Germany