Figure 2

a) The mean Ct of the housekeeping gene (GAPDH) is correlated with the 3' slope of mean log ratio of the signatures from FFPET muscle issues. The 3' slope of mean log ratio is one of the key quality metrics for microarray hybridization, which is the mean expression ratio of all genes as a function of distance to the 3' end of the transcript. Good quality samples that show no amplification bias will have a 3' slope of zero. b) One-dimensional heatmap of FFPET hybridization experiments starting with cRNAs of different qualities sorted by Ct count of the APPBP1. The Ct count for gene APPBP1 is a measure of the relative abundance of 'specific transcript' (amplifiable RNA) in the amplified cRNA for each sample. A distinctive gene expression pattern is correlated with the Ct count of cRNA for each individual hybridization. At p-value of 0.05, 4,500 genes are selected in this plot. The top panel shows the relative GC content for the corresponding signatures. c) Histogram showing the number of genes in each correlation between the mean Ct count of the amplified cRNA and the detected expression pattern. In an ideal hybridization, there should be no correlation between the Ct count of the APPBP1 for the amplified cRNA and the detected expression pattern, either positive or negative. However, the histogram of correlation coefficient with Ct counts for all genes detected on the FFPET array indicates that a significant fraction of genes is correlated or anti-correlated with the Ct count of the APPBP1 for the amplified cRNA, suggesting that the sample quality rather than the biology of FFPET samples dictates the correlation.