Figure 6
Comparing the reproducibility of different test statistics. Two subsets of Dataset 4 each contain three replicate arrays derived from identical biological experiments. Each test statistic is calculated twice, once for each subset, and the two statistics are plotted against each other. (a) Comparison of 90th percentile penalized statistics. (b) Comparison of SAM penalized statistics. (c) Comparison of Z-statistics using Iavg pooling metric. (d) Comparison of Z-statistics using Imin pooling metric. Also see Fig. 1a for comparison of the standard t-test.