Microarray performance metrics. (a,b) Frequency distribution of normalized microarray data. Histograms depict rat (a) and human (b) gene signals (rat: n = 36,685 genes; human: n = 53,617 genes) after data normalization and transformation (Log2). Each microarray dataset is represented by a distinct line. (c) Quality assessment metrics for each experimental group. Performance parameters before (background mean) and after (all other parameters) data normalization were monitored via Affymetrix Expression ConsoleTM software. Reported values are mean ± SD (n per experimental group = 3 microarrays). AUC, area under curve; MAD, mean absolute deviation; RLE, relative log expression (d,e) Correlation matrices of Pearson’s correlation coefficients. Intra- and inter-experimental group correlations were calculated based on Log2-transformed rat (d; n = 23,243 genes) and human (e; n = 33,424 genes) data. All correlations were highly significant (p < 0.001). Grey values are replicates of colour-coded correlation. (f) Average Pearson’s correlation coefficients of intra- and inter-experimental group comparisons. Colour codes are matched between panels d-f. Data are represented as mean ± SD. Intra-experimental group correlations: n = 3; inter-experimental group correlations: n = 9 (g,h) Dendrograms illustrating unsupervised hierarchical clustering (Pearson’s correlation; average linkage) of rat (g) and human (h) microarrays (rat: n = 23,243 genes; human: n = 33,424 genes).