A package designed to interactively explore biclusters
identified in gene expression data.
Introduction
Clustering genes according to their expression profiles is an important
task in analyzing microarray data. ExpressionView allows the user to
explore the biclusters together with the underlying data in an
interactive environment. The applet requires a browser with Adobe Flash
player 10 installed.
The ExpressionView package can treat gene expression data in the form of
a Bioconductor ExpressionSet. It recognizes
biclusters identified by the Iterative Signature Algorithm (ISA)
implemented in the ISA package as well as the
methods available in the biclust package.
Worflow
The usual workflow consist of three steps:
Order
To arrange the possibly overlapping biclusters in a visually
appealing layout, the gene expression data has to be reordered in such
a way that individual biclusters from contiguous rectangles. The
optimal order is found by the OrderEV function.
Export
In a second step, the biclusters and all the relevant
biological information are combined and exported to an ExpressionView
XML file. This is done by the ExportEV function.
Visualize
The Flash applet is started by the LaunchEV
command. Video tutorials describing the various features of the
applet can be found on the ExpressionView website
(http://www.unil.ch/cbg/ExpressionView).
A tutorial can be found in the accompanying vignette of the package.
Biclustering with non-gene expression data
Both, the ISA and the biclustering methods
implemented in the biclust package can treat any
two-dimensional data, i.e., not necessarily originating from gene
expression profiling. While the ExpressionView package is optimized for
gene expression matrices, it is also possible to explore data stemming
from other sources. For more information, see the description of the
ExportEV function.