From reference genes to global mean normalization

Abstract
The accuracy and precision of gene expression results are in large part determined by the applied normalization strategy to remove the experimentally induced variation in order to reveal the true biological changes. In this presentation, I will review state of the art normalization methods, going from classic reference genes through expressed repeat sequences to global mean normalization. Normalization against 3 or more validated reference genes is considered as the most appropriate and universally applicable method. While many algorithms have been reported to date to evaluate a handful of candidate reference genes in terms of expression stability (or suitability as normalizing gene) [1], the geNorm method [2] was the first (http://medgen.ugent.be/genorm) and has established itself as the de facto standard with more than 2000 citations (Google Scholar, December 2009). As an alternative to classic reference genes that require validation in each experimental system, we developed and validated a novel normalization method that is based on expressed repeat sequences. Using normal and diseased human and mouse samples, we could demonstrate that various repeat elements are expressed and can be safely used as surrogate marker for the mRNA content of the sample. As such, an RT-qPCR assay that targets these repeats enables quick and validation-free normalization [3]. Finally, when measuring a large unbiased set of genes, another straightforward and universal method can be used to normalize gene expression results. Using microRNA expression profiles from different human tissues as a model system, we could demonstrate that the mean expression level of all expressed microRNA genes constitutes the best normalization factor. It removes most of the experimentally induced variation and is able to better appreciate small biological changes [4].