DNA microarrays allow the simultaneous measurement of gene expression levels for a large number of genes in several experimental samples. Cluster analysis applied to microarray data has been proposed to discover previously unknown sets of co-regulated genes with similar expression profiles across samples, perhaps reflecting similar gene function or participation in the same biological pathway. I will describe recent results on the plaid model, a new two-way clustering method motivated by microarray data. The plaid model is designed to identify gene groups in which similarity extends across only a subset of the experimental samples. Within a cluster, the model allows a distinct regulatory effect for each gene and each sample. Genes and samples can belong to one, more than one or none of the groupings identified in the data. This research is joint work with Art Owen.
Copyright. All Rights Reserved.