Figure 5From: A power law global error model for the identification of differentially expressed genes in microarray dataComparison of two methods for inferring the null distribution of the PLGEM-STN statistic. The classic permutation strategy (excluding DEG) was performed for each comparison in the Latin Square data set and the quantiles of the distribution of PLGEM-STN values were averaged over the 13 comparisons. The mean quantiles of the permutated statistics are plotted against the quantiles of the distribution of PLGEM-STN values obtained through the proposed resampling approach performed on the same data set but including DEG.Back to article page