Principle components analysis (PCA) and sources of variation. Principle components analysis (PCA) of 71 samples displaying spatial separation (expression level clustering) of RNA extracted from FF samples (shown in red on left) and FFPE samples (shown in blue on right); The three principle components (PC#1, #2 & #3 and respective contributable variations are presented on x-, y- and z-axis along). A Statistical significance of the different "sources of variation" in the total gene expression estimated by 3-way ANOVA models. F ratio for each factor (source) represents the F-statistics for that factor/F-statistics for error (noise). The mean of the F Ratio for all the genes for a particular factor are shown on Y-axis and the different factors (sources of variation) are shown in x-axis.