PCA and ANOVA of Transcript Expression Data. Analysis of Variance (ANOVA) was performed on the entire data set and a gene list was generated Figure 2a shows the histogram of variance indicating that the major variance was due to the differences between tumor and normal samples and scan date variance is almost the same as that attributed to type 1 error. PCA was used reduce dimensionality and to examine whether clusters of tumors separated from normal samples or if other variables, such as scan date contributed to the variance. It can be seen that the tumors are spread throughout the three dimensional space of the plot while the normal samples form a tight cluster at the base of the the plot on the right. Concordance of the clustering structure was observed between the two dimensionality-reduction procedures (data not shown). Analysis of Variance (ANOVA) was then performed on the entire data set and a gene list was generated using an FDR of 0.05, 2-fold expression changes and a p-value for fold change of 0.05. Figure 2 shows the histogram of variance and the PCA plot of the effect of the scan date on the data.