Figure 1

Example of a comparative microarray experiment (treated versus untreated or control). The resulting intensity data are filtered non-specifically to reduce the number of hypotheses to be tested. In the next step, relative expression values and p-values are calculated for every single gene. After geneset reduction, several sophisticated methods for gene subset selection can be applied. Principal component analysis (PCA) is mainly used for dimensionality reduction. PCA enables the visualization of multidimensional datasets and can be effectively applied for gene selection. Other approaches for subset selection are functional enrichment strategies. These methods result in sets of aggregating genes with similar functions. Cluster analysis can be used to group genes according to their expression similarity. Clusters of genes with similar expression profiles can be used as starting points for further bioinformatic analyses. Network analysis is a method where an interaction network is constructed by integrating the geneset with direct or indirect molecular relationships extracted from various biological knowledge bases. The subnetworks show a distinct degree of interconnectivity. For example, subnetwork 7 appears to contain interactions that are responsible for some important process or behaviour, as its removal would affect the entire network. (Herein, network images were created via Ingenuity^® Pathway Analysis (IPA); http://www.ingenuity.com).