Table 2 Illustration of the empirical Power based on a nominal two-sided 0.05 FWER level, with m= 200 genes, N= 200 simulation samples and B= 200 resampling replications per sample. The outlier law realizes values -4,0 or 4 with probabilities π/2, 1 - π and π/2 respectively. Under simulation model 1, the outlier effect is added to all time-points. Under simulation model 2, the outlier effect is only added to the first and last time-points.
From: Robust test method for time-course microarray experiments
Simulation 1 | Simulation 2 | ||||||
|---|---|---|---|---|---|---|---|
π | ρ |
|
|
|
|
|
|
0 | 0 | 0.960 | 0.995 | 0.995 | |||
0.3 | 0.880 | 0.940 | 0.925 | ||||
0.6 | 0.750 | 0.855 | 0.845 | ||||
0.05 | 0 | 0.910 | 0.800 | 0.790 | 0.890 | 0.880 | 0.970 |
0.3 | 0.835 | 0.755 | 0.755 | 0.845 | 0.825 | 0.900 | |
0.6 | 0.680 | 0.630 | 0.605 | 0.835 | 0.805 | 0.935 | |
0.1 | 0 | 0.760 | 0.485 | 0.480 | 0.895 | 0.740 | 0.930 |
0.3 | 0.675 | 0.415 | 0.425 | 0.660 | 0.585 | 0.885 | |
0.6 | 0.650 | 0.410 | 0.370 | 0.730 | 0.530 | 0.895 | |
-PERM
-PERM