Table 2.
Comparison of RMA and fRMA based on bias, precision, and overall performance
| Preprocessing | Slope (SD) | Null SD | Null 99.5% | SNR | POT | |
|---|---|---|---|---|---|---|
| Low | RMA | 0.14 (0.54) | 0.47 | 1.46 | 0.30 | 0.01 |
| fRMA | 0.20 (0.48) | 0.40 | 1.24 | 0.50 | 0.02 | |
| Medium | RMA | 0.69 (0.55) | 0.39 | 1.26 | 1.77 | 0.15 |
| fRMA | 0.75 (0.49) | 0.34 | 1.11 | 2.21 | 0.23 | |
| High | RMA | 0.61 (0.42) | 0.33 | 0.99 | 1.85 | 0.18 |
| fRMA | 0.60 (0.36) | 0.28 | 0.86 | 2.14 | 0.23 | |
Comparison of RMA and fRMA trained on the modified spike-in data. For three intensity strata, we report assessments of accuracy (column 1), precision (columns 2 & 3), and overall performance (columns 4 & 5).