Table 3.
Detection of variants from the STMM-Mix-II control DNA mixtures
| DNA Input (ng) | Expected variant allelic fraction (VAF) | Calls | Accuracy | ||||
|---|---|---|---|---|---|---|---|
| FN | TP | FP | Sensitivity | Specificity | PPV | ||
| 100 | 40% | 0 | 36 | 1 | 100.0% | 100.0% | 97.3% |
| 25% | 1 | 35 | 2 | 97.2% | 100.00% | 94.6% | |
| 15% | 0 | 36 | 6 | 100.0% | 100.00% | 85.7% | |
| 10% | 1 | 35 | 2 | 97.2% | 100.00% | 94.6% | |
| 5% | 0 | 36 | 5 | 100.0% | 100.00% | 87.8% | |
Analysis includes 37 variants of the Seraseq STMM-II that overlap regions with sufficient read coverage. Analysis was performed with the Compass analysis software (TOMA Biosystems Inc., Foster City, CA), removing PCR duplicates. TP true positives, the variants in the call set that match the variants in ground truth list for the reference material provided by the manufacturer; FN false negatives, the variants that are in the ground truth list, but not present in the call set; FP false positives, variants in the call set that are not present in the ground truth list. The ground truth list only includes the spiked-in synthetic somatic mutations and not the germline variants present in the background genome