MOTIVATION: Microarray technology has become a powerful research tool in many fields of study; however, the cost of microarrays often results in the use of a low number of replicates (k). Under circumstances where k is low, it becomes difficult to perform standard statistical tests to extract the most biologically significant experimental results. Other more advanced statistical tests have been developed; however, their use and interpretation often remain difficult to implement in routine biological research. The present work outlines a method that achieves sufficient statistical power for selecting differentially expressed genes under conditions of low k, while remaining as an intuitive and computationally efficient procedure. RESULTS: The present study describes a Global Error Assessment (GEA) methodology to select differentially expressed genes in microarray datasets, and was developed using an in vitro experiment that compared control and interferon-gamma treated skin cells. In this experiment, up to nine replicates were used to confidently estimate error, thereby enabling methods of different statistical power to be compared. Gene expression results of a similar absolute expression are binned, so as to enable a highly accurate local estimate of the mean squared error within conditions. The model then relates variability of gene expression in each bin to absolute expression levels and uses this in a test derived from the classical ANOVA. The GEA selection method is compared with both the classical and permutational ANOVA tests, and demonstrates an increased stability, robustness and confidence in gene selection. A subset of the selected genes were validated by real-time reverse transcription-polymerase chain reaction (RT-PCR). All these results suggest that GEA methodology is (i) suitable for selection of differentially expressed genes in microarray data, (ii) intuitive and computationally efficient and (iii) especially advantageous under conditions of low k. AVAILABILITY: The GEA code for R software is freely available upon request to authors.

MOTIVATION: A crucial step in microarray data analysis is the selection of subsets of interesting genes from the initial set of genes. In many cases, especially when comparing a specific condition to a reference, the genes of interest are those which are differentially expressed. Two common methods for gene selection are: (a) selection by fold difference (at least n fold variation) and (b) selection by altered ratio (at least n standard deviations away from the mean ratio). RESULTS: The novel method proposed here is based on ANOVA and uses replicate spots to estimate an empirical distribution of the noise. The measured intensity range is divided in a number of intervals. A noise distribution is constructed for each such interval. Bootstrapping is used to map the desired confidence levels from the noise distribution corresponding to a given interval to the measured log ratios in that interval. If the method is applied on individual arrays having replicate spots, the method can calculate an overall width of the noise distribution which can be used as an indicator of the array quality. We compared this method with the fold change and unusual ratio method. We also discuss the relationship with an ANOVA model proposed by Churchill et al. In silico experiments were performed while controlling the degree of regulation as well as the amount of noise. Such experiments show the performance of the classical methods can be very unsatisfactory. We also compared the results of the 2-fold method with the results of the noise sampling method using pre and post immortalization cell lines derived from the MDAH041 fibroblasts hybridized on Affymetrix GeneChip arrays. The 2-fold method reported 198 genes as upregulated and 493 genes as downregulated. The noise sampling method reported 98 gene upregulated and 240 genes downregulated at the 99.99% confidence level. The methods agreed on 221 genes downregulated and 66 genes upregulated. Fourteen genes from the subset of genes reported by both methods were all confirmed by Q-RT-PCR. Alternative assays on various subsets of genes on which the two methods disagreed suggested that the noise sampling method is likely to provide fewer false positives.