PARIS activity center analysis

Following are instructions for using the PARIS activity center analysis web application.

The application compares two samples chosen from a GEO microarray sample series, displaying functional neighborhood "activity" values graphically.

The functional network is for human genes only, so choose sample series that are annotated as coming from human tissue samples or cell cultures.

Thanks to Millennium Pharmaceuticals, Inc. for providing the code under a BSD license.

Install Firefox 3
Firefox is needed in order to run Greasemonkey.

Although the graphical results viewer will run in Firefox 2, results are better in Firefox 3, which has an improved SVG viewer.

Install Greasemonkey
Greasemonkey is a Firefox plugin that enables the use of "user scripts" that manipulate loaded web pages. Get it from https://addons.mozilla.org/en-US/firefox/addon/748. More information at http://userscripts.org/.

Install sample selection script (uses Greasemonkey)
From http://userscripts.org/scripts/show/12212

Go to GEO and select samples
Run Firefox and go to NCBI GEO (Gene Expression Omnibus) and find a data set of interest. Data sets have accession numbers starting with "GDS".

Select a "before" or "control" sample by clicking on one of the yellow "+A" boxes.

Select an "after" or "experimental" sample by clicking on one of the blue "+B" boxes.

The main search page for GEO data sets is http://www.ncbi.nlm.nih.gov/geo/

Example: http://www.ncbi.nlm.nih.gov/sites/entrez?db=gds&term=GPL96%5BAccession%5D&cmd=search

Sample combination is not yet implemented, so for now please select only one 'A' and one 'B' sample.

Go to analysis application start page
After selecting A and B samples, click on the "Compare!" button.

This will take you to a page showing you which GEO samples you have selected for your A and B sets. Example

If you know the identifiers for the data set and the A and B samples you want to compare, you can go directly here without using Greasemonkey.

Perform analysis and view "cobweb" plot
Click on 'Display results as SVG' and wait a while (maybe 30 seconds).