Table 2. Evaluation of SMD somatic SNV detection in the simulation datasets.
| Variant Set | Dataset 1 | Dataset 2 | ||||||
|---|---|---|---|---|---|---|---|---|
| TP | FN | Sensitivity % | FNR | FP | TN | Specificity % | FPR | |
| BWA-mem+MuTect | 690 | 52 | 92.99 | 7.01E-02 | 16 | 47,301,677 | 99.99997 | 3.38E-07 |
| Novoalign+MuTect | 684 | 58 | 92.18 | 7.82E-02 | 23 | 47,301,670 | 99.99995 | 4.86E-07 |
| BWA-mem+Shimmer | 550 | 192 | 74.12 | 2.59E-01 | 0 | 47,301,693 | 100.00000 | 0.00 |
| Novoalign+Shimmer | 536 | 206 | 72.24 | 2.78E-01 | 0 | 47,301,693 | 100.00000 | 0.00 |
| BWA-mem+SomaticSniper | 707 | 35 | 95.28 | 4.72E-02 | 110 | 47,301,583 | 99.99977 | 2.33E-06 |
| Novoalign+SomaticSniper | 697 | 45 | 93.94 | 6.06E-02 | 109 | 47,301,584 | 99.99977 | 2.30E-06 |
| BWA-mem+Strelka | 597 | 145 | 80.46 | 1.95E-01 | 16 | 47,301,677 | 99.99997 | 3.38E-07 |
| Novoalign+Strelka | 596 | 146 | 80.32 | 1.97E-01 | 19 | 47,301,674 | 99.99996 | 4.02E-07 |
| BWA-mem+VarScan2 | 708 | 34 | 95.42 | 4.58E-02 | 27 | 47,301,666 | 99.99994 | 5.71E-07 |
| Novoalign+VarScan2 | 705 | 37 | 95.01 | 4.99E-02 | 25 | 47,301,668 | 99.99995 | 5.29E-07 |
| BWA-mem+Virmid | 678 | 64 | 91.37 | 8.63E-02 | 0 | 47,301,693 | 100.00000 | 0.00 |
| Novoalign+Virmid | 690 | 52 | 92.99 | 7.01E-02 | 4 | 47,301,689 | 99.99999 | 8.46E-08 |
| 2aligners+2callers | 713 | 29 | 96.09 | 3.91E-02 | 1 | 47,301,692 | 100.00000 | 0.00 |
Results are shown for high-quality variants that passed all quality filters. Additional precision digits were kept for Specificity to infer small differences.