Table 2 Accuracy, sensitivity, specificity, and PPV for benchmark datasets with simulated 454-style errors
From: Short-read reading-frame predictors are not created equal: sequence error causes loss of signal
Overall reading-frame accuracy | |||||
|---|---|---|---|---|---|
Dataset | FGS | MGA | MGM | OPH | PRODG |
0.00% | 91.0% | 93.6% | 94.5% | 90.5% | 91.7% |
0.20% | 87.8% | 81.3% | 82.3% | 76.0% | 78.9% |
0.50% | 83.8% | 69.7% | 70.5% | 62.9% | 66.1% |
2.80% | 58.4% | 25.9% | 26.0% | 22.1% | 23.2% |
Sensitivity | |||||
0.00% | 95.2% | 94.7% | 95.9% | 92.8% | 95.0% |
0.20% | 91.5% | 80.8% | 82.1% | 76.3% | 80.7% |
0.50% | 87.1% | 67.7% | 68.6% | 61.3% | 66.5% |
2.80% | 59.7% | 18.3% | 18.2% | 15.0% | 19.4% |
Specificity | |||||
0.00% | 59.0% | 84.2% | 82.9% | 71.8% | 68.5% |
0.20% | 59.7% | 85.1% | 84.0% | 73.3% | 66.7% |
0.50% | 58.8% | 85.9% | 85.6% | 75.5% | 66.4% |
2.80% | 49.0% | 89.2% | 89.8% | 81.5% | 65.8% |
Positive Predictive Value | |||||
0.00% | 91.6% | 96.1% | 96.4% | 93.2% | 94.0% |
0.20% | 88.9% | 90.9% | 91.2% | 86.5% | 86.6% |
0.50% | 85.5% | 86.0% | 85.9% | 80.2% | 78.1% |
2.80% | 62.1% | 58.0% | 56.5% | 44.0% | 35.4% |