Region-based custom chip description formats for reanalysis of publicly available affymetrix® genechip® data sets

Abstract

Commercially developed microarrays, such as those from Agilent® and Affymetrix®, allow for the analysis of differential gene expression changes on a genome-wide scale. Publicly repositories of microarray data, most notably ArrayExpress and the Gene Expression Omnibus (GEO) have made available millions of microarray samples to researchers worldwide. One of the drawbacks of microarray technology is the static construction of probes based on current genomic knowledge and gene annotation information available at the design phase. As the knowledge base about genes expands, including alternative isoform formation and alternative polyadenylation signaling, the need for a dynamically changing approach to microarray expression analysis has become apparent. We have therefore designed a framework for the reanalysis of publicly available microarray datasets by updating probe set construction based on gene, transcript, and region-based (UTR, exon, CDS) annotations. Our analysis of two publicly available GEO series, GSE48611 and GSE72551, illustrate that the analysis of expression changes using different annotation groupings yields additional insight into changes in transcript expression, in particular, 3' UTR dynamics, which are likely to present phenotypical differences.