Table 4.
Posteror Peptide Inference Improves the Peptide Identification
| |
|
PP |
BI |
||||
|---|---|---|---|---|---|---|---|
| Experiment | No. of true proteins | AUC | acc | F-measure | AUC | acc | F-measure |
| 1 | 46 | 0.953 | 0.896 | 0.884 | 0.996 | 0.975 | 0.973 |
| 2 | 46 | 0.953 | 0.896 | 0.886 | 0.997 | 0.972 | 0.960 |
| 3 | 46 | 0.974 | 0.927 | 0.926 | 0.999 | 0.966 | 0.964 |
The Area Under the ROC Curve (AUC), the accuracy (acc) and the F-measure are used to compare the peptide identification results from PeptideProphet (PP) and after Bayesian inference (BI). The peptides identified from the model and contaminant proteins were considered as positive, whereas the rest identified peptides were considered negative.