Skip to main content
Figure 7 | BMC Bioinformatics

Figure 7

From: Preferred analysis methods for Affymetrix GeneChips. II. An expanded, balanced, wholly-defined spike-in dataset

Figure 7

Accuracy of false discovery rate estimation on the Platinum Spike dataset. (A) The accuracy of the predicted FDR of the top 10 routes based on 1500 permutations (left), or the method of Benjamini and Hochberg (right). The predicted FDRs for the latter are only available for routes that report p-values. (B) The p-value distribution of null probe sets for the routes ranked as the first (left) and eighth (right) respectively to present two distinct distribution patterns observed for five of the top 10 routes whose p-values are available. "Null" probe sets includes both probe sets present with fold change = 1 and empty probe sets. Inset shows the p-value distribution of present probe sets with fold change = 1 only.

Back to article page