Figure 1

Scattersmooth plots of the correlation between the ranks (according to p values) of genes in the UG dataset of the 5 platforms. Red corresponds to denser areas, while yellow corresponds to non dense areas. The scattersmooth uses an algorithm for smoothing of two dimensional histograms with smoothed densities (26). This graph is more meaningful than a traditional scatter plot of the p values or of the -log p values, where the smallnumber of DEGs in our datasets originates graph blurred with thousands of overlapping dots and empty areas. Since the different signal to noise ratio is varying in the platforms and affects the statistics differently, plots of the ranks are more meaningful than plot of p values and statistics.