ABSTRACT
While finding genes that differ between 2 or more classes of samples is frequently the
first question addressed in the context of microarray studies, there are other questions
of interest. Most of these further issues have to do with examination of the manner in
which sets of genes work in concert to perform complex physiological phenomena.
For example, multiple genes are typically involved to respond to some extra-cellular
signal, hence if a cell receives a signal then all genes involved in the response should
have their expression level change. This would induce correlations in the level of
gene expression for all genes involved in this cellular response. In addition to finding
groups of genes that work together, we may also seek sets of subjects that seem to
belong to the same group in terms of their gene expression profiles (i.e. the set of
gene expression measurements for all genes). Some methods seek to perform both of
these task simultaneously. For concreteness, most of the discussion here will focus
on finding groups of genes that behave in similar ways in response to a variety of
stimuli.