Table 11.
Definitions used for metrics calculations
| Metric | Description | Function |
|---|---|---|
| True Positive (TP) | Result is positive, key is positive | |
| True Negative (TN) | Result is negative, key is negative | |
| False Positive (FP) | Result is positive, key is negative | |
| False Negative (FN) | Result is negative, key is positive | |
| Precision (PPV) | Positive Predictive Value; number of TP of result positives | TP/(TP + FP) |
| Recall (TPR) | Sensitivity; number of TP out of all key positives | TP/(TP + FN) |
| Specificity (TNR) | Number of TN out of all key negatives | TN/(FP + TN) |
| False Positive Rate (FPR) | Number of false positives out of all key negatives | FP/(FP + TN) |
Each sample from real-time PCR and sequencing was measured for a target assay (columns faceted along top of Figs. 7, 8, 9, 10). Under Description, ‘Result’ is in reference to a real-time PCR Ct value or a sequence mapped count above (negative for PCR, positive for sequencing) or below (positive for PCR, negative for sequencing) a particular threshold. These definitions were used to generate the Precision-recall (PR) and Receiver Operator Characteristic (ROC) curves in Figs. 7, 8, 9, 10. A range of thresholds was employed for the generation of PR and ROC curves