Skip to main content
Figure 3 | BMC Bioinformatics

Figure 3

From: Probe-level linear model fitting and mixture modeling results in high accuracy detection of differential gene expression

Figure 3

ROC curves comparing methods on the Choe et al. dataset. Probe-sets with a spiked to control concentration ratio above or equal to 1.2 were considered as DEGs to compute these curves, resulting in 1,326 spiked probe-sets to identify among a total of 14,010. The fold-change applied after RMA summaries and quantile normalization is shown as a baseline for comparison since it is a simple and frequently used method. Nine shaded curves, representing the results from applying the PL-LM method to all combinations of two arrays (one control vs. one spike), are falling directly under the PL-LM (on all replicates) curve (black line). For all methods, the area under the curve (AUC) is reported as a quantitative measure of both sensitivity and specificity.

Back to article page