Abstract
DNA microarrays are a powerful tool to investigate differential gene expression for thousands of genes simultaneously. Although DNA microarrays have been widely used to understand the critical events underlying growth, development, homeostasis, behavior and the onset of disease, the management of the resulting data has received little attention. Presently, the fluorescent dyes Cy3 and Cy5 are most often used to prepare labeled cDNA for microarray hybridizations. Raw microarray data are image files that have to be transformed into gene expression formats a process that requires data manipulation due to systematic variations which may be attributed to differences in the physical and chemical dye characteristics. Since the goal of most microarray applications is to identify differences in transcript levels calculated from fluorescence ratios it is necessary to normalize fluorescence signals to compensate for systematic variations. Here, we will review current normalization strategies applied to cDNA microarrays and discuss their limits. We will show that experimental design determines normalization success.