ABSTRACT
Normalization of microarray data is any procedure meant to reduce or account for systematic variation among or within arrays. This variation is a component common to all the genes that are measured on an array or, more generally, to a subset of genes on the array. Normalization methods are often applied prior to the application of statistical analysis methods, which are usually designed to detect differential expression. However, normalization includes procedures that adjust for known effects as part of the statistical analysis as well as those that replace the actual data with modified values prior to the statistical analysis. In either case, normalization represents an effort to obtain more powerful tests by reducing variation in the data or otherwise accounting for it mathematically.
