Data Normalization and analysis
Normalization is necessary to adjust for differences in labeling and detection efficiencies of the fluorescent labels and for differences in the quantity of starting RNA. Data was normalized using a local regression technique LOWESS (LOcally WEighted Scatterplot Smoothing) for hybridizations with RNA-based samples using a software tool MIDAS ([43], TIGR, Rockville, MD), while total intensity normalization was used for the hybridizations with genomic DNA samples. The resulting data was averaged from triplicate genes on each array and from duplicate flip-dye arrays for each experiment.
Differentially expressed genes at the 95% confidence level were determined using intensity-dependent Z-scores (with Z = 1.96) as implemented in MIDAS. The resulting lists of the genes were examined further by cross comparison between experiments using TIGR MEV [43], TIGR, Rockville, MD) using Euclidean distance and hierarchical clustering with average linkage clustering method.