Table 2. CNV prediction results for gold standard datasets.
Results indicate 10%/50% reciprocal overlap between predicted CNV and gold standard. CNV predictions not meeting overlap criteria were classified as “Other.”
| Deletion | Duplication | |||||
|---|---|---|---|---|---|---|
| Algorithm | Sensitivity | True Positives | Other | Sensitivity | True Positives | Other |
| NA12878 (low coverage) | ||||||
| CNVnator | 0.21 / 0.16 | 102 / 78 | 548 / 573 | 0.14 / 0.08 | 28 / 17 | 206 / 218 |
| RDXplorer | 0.07 / 0.05 | 37 / 23 | 218 / 234 | 0.03 / 0.01 | 7 / 3 | 349 / 355 |
| GROM-RD | 0.44 / 0.37 | 217 / 181 | 863 / 901 | 0.15 / 0.11 | 31 / 22 | 313 / 322 |
| NA12878 (high coverage) | ||||||
| CNVnator | 0.79 / 0.68 | 391 / 341 | 27597 / 27653 | 0.15 / 0.10 | 34 / 23 | 975 / 989 |
| RDXplorer | 0.23 / 0.18 | 117 / 92 | 1650 / 1679 | 0.10 / 0.05 | 22 / 12 | 794 / 806 |
| GROM-RD | 0.71 / 0.61 | 352 / 303 | 5395 / 5438 | 0.20 / 0.15 | 45 / 34 | 1464 / 1472 |