Skip to main content

Table 4 Percentage in decrease of mean FDR

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

  ρ True π0 BYa SampG.qb SampS.qc SampG.md SampS.me
Data-B 0.3 0.9 84.13% 2.19% 3.00% 1.99% 2.05%
   0.8 90.03% 5.54% 5.60% 4.04% 4.33%
   0.7 87.04% 7.64% 8.27% 7.44% 7.84%
   0.6 91.59% 6.82% 6.72% 7.09% 7.45%
  0.5 0.9 92.89% 4.51% 8.51% 1.83% 6.39%
   0.8 92.87% 7.68% 9.73% 3.90% 7.66%
   0.7 95.09% 6.79% 11.63% 3.01% 8.24%
   0.6 94.24% 8.49% 13.26% 4.34% 8.82%
  0.7 0.9 96.62% 19.79% 35.22% 18.30% 25.52%
   0.8 97.03% 28.32% 40.86% 26.00% 32.72%
   0.7 96.15% 17.09% 35.83% 14.98% 23.57%
   0.6 97.80% 20.12% 36.47% 15.41% 27.46%
Data-M   0.9 94.87% 10.25% 14.66% 7.00% 14.70%
   0.8 93.11% 7.87% 19.48% 6.33% 11.38%
   0.7 93.50% 4.60% 12.11% 0.15% 5.96%
   0.6 94.18% 6.35% 14.16% 1.85% 8.70%
  1. a (FDRqvalue - FDRBY)/FDRqvalue
  2. b (FDRqvalue - FDRSampG.q)/FDRqvalue
  3. c (FDRqvalue - FDRSampS.q)/FDRqvalue
  4. d (FDRqvalue - FDRSampG.m)/FDRqvalue
  5. e (FDRqvalue - FDRSampS.m)/FDRqvalue