Skip to main content
Figure 5 | BMC Bioinformatics

Figure 5

From: Re-sampling strategy to improve the estimation of number of null hypotheses in FDR control under strong correlation structures

Figure 5

Boxplot of π 0 estimated by various methods on the simulated data sets. The methods are (from left to right in each figure): LSL, Qvalue, SPLOSH, BUM, convest, PRE, mgf, adaptive, SampG.q Q2, SampG.q Q3, SampS.q Q2, SampS.q Q3, SampG.m Q2, SampG.m Q3, SampS.m Q2, and SampS.m Q3. A-D: Data-B with ρ = 0.3 and π0 varies from 0.9 to 0.6. E-H: Data-B with ρ = 0.5 and π0 varies from 0.9 to 0.6. I-L: Data-B with ρ = 0.7 and π0 varies from 0.9 to 0.6. M-P: Data-M with π0 varies from 0.9 to 0.6. Horizontal lines represent the true π0 in each simulated case.

Back to article page