Position-based clustering of microarray expression data

Christopher Faulk, Louisiana State University
Joomyeong Kim, Louisiana State University


Microarray expression analysis has traditionally focused only on genes with the highest expression level changes, ignoring the majority of genes with lower fold changes. To address this problem, we provide a simple method that can derive additional useful biological information from the rest of the genes included on an array. Chromosomal position information combined with expression data, either raw or in expression-level changes, gives a regional overview of expression from an array. This protocol is tailored to Affymetrix data, but can be used for other types of microarray results. The procedure is illustrated by the reanalysis of a data set, GSE5230, previously deposited in the Gene Expression Omnibus (GEO). Using this procedure, we identified several classes of chromosomal regions where expression levels were affected in concert. Linking expression data to chromosomal position also allowed us to identify several genomic regions displaying low but steady fold change, which were missed by traditional approaches. Overall, this method is useful in detecting regional regulatory changes. It should allow for greater use of the large quantity of previously overlooked microarray data. Copyright © 2009 by Cold Spring Harbor Laboratory Press.