SAM Threshold and False Discovery Rate
| SAM threshold | Original data | Permuted data | Difference | FDR |
| 0.10 | 13,113 | 3757 | 9356 | 28.6% |
| 0.13 | 10,627 | 1894 | 8733 | 17.8% |
| 0.15 | 9191 | 1300 | 7891 | 14.1% |
| 0.20 | 7088 | 718 | 6370 | 10.1% |
| 0.30 | 4837 | 294 | 4543 | 6.1% |
| 0.40 | 3806 | 179 | 3627 | 4.7% |
| 0.50 | 3170 | 129 | 3041 | 4.1% |
[i] The number of features exceeding a given threshold in the original data and the average number of features exceeding the threshold in the permuted data set is shown. This is used to calculate the difference and the false discovery rate (FDR) at different thresholds according to SAM (Tusher et al. 2001).