ABSTRACT
Expression profiling is by far the most common application of microarray technology. For the majority of model organisms there are microarrays available that represent the entire transcriptional capacity of a cell. On the other hand the relatively low cost of manufacturing a large collection of PCR amplified cDNA clones, or of acquiring an oligonucleotide set designed to the organism of interest makes possible the development of microarrays that have good genome coverage for a large number of non model organisms. As a consequence huge volumes of transcriptomics data have become available in publicly accessible databases such as Gene Expression Omnibus (Barret et al., 2005) or Array Express (https://www.ebi.ac.uk/arrayexpress/). Experimental design and the data analysis strategy are the two single most important factors in the success of an expression profiling experiment. In this chapter we describe two experiments that have been performed in our laboratory to characterize two different aspects of bacterial pathogenesis. These two case studies represent both prokaryotic (the adaptation of Escherichia coli cells to temperature shift between 10°C and 37°C) and eukaryotic (the response of host cells to bacterial infection) systems and use two different experimental designs (single-and two-channel microarrays). Although the data analysis techniques used to analyze expression profiling data are usually applicable to both types of microarrays, data processing, especially the normalization procedures, differ substantially.
