Abstract Motivation: When analyzing microarray data, non-biological variation introduces uncertainty in the analysis and interpretation. In this paper we focus on the validation of significant differences in gene expression levels, or normalized channel intensity levels with respect to different experimental conditions and with replicated measurements. A myriad of methods have been proposed to study differences in gene expression levels and to assign significance values as a measure of confidence. In this paper we compare several methods, including SAM, regularized t-test, mixture modeling, Wilk's lambda score and variance stabilization. From this comparison we developed a weighted resampling approach and applied it to gene deletions in Mycobacterium bovis. Results: We discuss the assumptions, model structure, computational complexity and applicability to microarray data. The results of our study justified the theoretical basis of the weighted resampling approach, which clearly outperforms the others. Availability: Algorithms were implemented using the statistical programming language R and available on the author's web-page. Supplementary information: For additional material see http://www.sbi.uni-rostock.de/

MOTIVATION: Microarray experiments are inherently noisy. Replication is the key to estimating realistic fold-changes despite such noise. In the analysis of the various sources of noise the dependency structure of the replication needs to be taken into account. RESULTS: We analyzed replicate data sets from a Mycobacterium tuberculosis trcS mutant in order to identify differentially expressed genes and suggest new methods for filtering and normalizing raw array data and for imputing missing values. Mixed ANOVA models are applied to quantify the various sources of error. Such analysis also allows us to determine the optimal number of samples and arrays. Significance values for differential expression are obtained by a hierarchical bootstrapping scheme on scaled residuals. Four highly upregulated genes, including bfrB, were analyzed further. We observed an artefact, where transcriptional readthrough from these genes led to apparent upregulation of adjacent genes. AVAILABILITY: All methods and data discussed are available in the package YASMAhttp://www.cryst.bbk.ac.uk/wernisch/yasma.html for the statistical data analysis system R (http://www.R-project.org).