Help with Interpreting Gene Expression Data

Home
Return to the Homepage

Explore
Interactively explore the data

Figures
View figures from the paper

Download
Download the primary data

Help
Get help with interpretation of the data

People
Find Individuals & Institutions involved in this project

Help:

How do I interpret the cluster images?
Hierarchical clustering arranges genes according to their similarity in expression profiles across all of the array experiments, such that genes with similar expression patterns are "clustered" together. The gene expression data are displayed in tabular format in which each row of colored boxes corresponds to the variation in transcript level for each gene, and each column represents the variation in transcript levels of every gene in a given mRNA sample, as measured on one array. The variation in transcript abundance is represented by a color scale in which shades of red represent increases and shades of green decreases in transcript abundance, relative to the unstressed reference sample, and the saturation of the color corresponds to the magnitude of change. A black color indicates an undetectable change in transcript abundance, and a grey color indicates missing data.

How is the similarity between two genes determined?
The similarity of two genes in expression space is determined by calculating a distance measure across all experimental values. For this we are using the Pearson correlation coefficient.

Problems viewing images
We recommend that you upgrade to the latest browser versions. For Netscape this is version 4.51 for PowerPC Macintoshes, and Windows95/98 PCs. For 68K Macintoshes, and Windows 3.1 PCs you should get Netscape version 4.08. If you use Internet Explorer you should use version 4.5 on a Mac, and version 5.0 on a PC. If you still have problems then it may be that your computer doesn't have enough memory, or a fast enough processor. There are PDF versions of all figures which are somewhat less resource intensive.