| One-step analysis | Two-step analysis |
---|
# Selected
|
# Target
|
Est FDR
|
True FDR
|
# Target
|
Est FDR
|
True FDR
|
100
| 75 | 49 | 25 | 88 | 17 | 12 |
200
| 133 | 58 | 34 | 167 | 23 | 17 |
300
| 183 | 67 | 39 | 238 | 29 | 21 |
400
| 229 | 75 | 43 | 298 | 35 | 26 |
500
| 272 | 80 | 47 | 336 | 44 | 33 |
- Two groups of data were compared under two conditions. The first group is composed of two synthetic data sets made from the experimental replicates Wt-t0a and Wt-t0b, in which the same subset of 500 genes is increased (p <= 0.10), and the second group is composed of two synthetic data sets made from the same experimental replicates, in which no gene was changed. The one-step analysis consists of processing all of the variations together after the standardization step, leading to the generation of only one F curve, whereas with the two-step analysis the observed variation distributions of increased and decreased genes are considered independently, with each having its own F curve (Finc and Fdec). FDR was chosen so as to obtain a predetermined number of selected genes indicated in the first column. The second and fifth columns (# Target) give the number of true positivesselected. The third and sixth column (Est FDR) indicate the level of selection applied and give an estimate of the FDR. The fourth and seventh columns give the true FDR value, corresponding to the ratio of the number of true positives to the number of genes selected.