Table 3.
Large deletions and insertions (≥50 bp) calling performance across three CLR data sets
| Library | Metric | FocalSV (target) | FocalSV (auto) | PAV | SVIM-asm | Dipcall | cuteSV | SVIM | pbsv | Sniffles2 | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| DEL | CLR_L1 | Precision | 91.76% | 93.32% | 81.51% | 88.51% | 82.40% | 92.04% | 91.80% | 93.99% | 92.49% |
| Recall | 94.17% | 91.59% | 86.88% | 86.10% | 21.04% | 89.38% | 91.70% | 91.89% | 91.81% | ||
| F1 | 92.95% | 92.45% | 84.11% | 87.29% | 33.52% | 90.69% | 91.70% | 92.92% | 92.15% | ||
| GT concordance | 98.76% | 98.73% | 96.00% | 95.23% | 85.45% | 98.64% | 97.90% | 96.64% | 98.68% | ||
| CLR_L2 | Precision | 92.03% | 93.43% | 91.21% | 92.01% | 88.81% | 91.43% | 93.40% | 93.29% | 92.93% | |
| Recall | 94.83% | 93.97% | 89.02% | 91.72% | 86.95% | 91.98% | 93.90% | 92.95% | 92.23% | ||
| F1 | 93.41% | 93.70% | 90.10% | 91.86% | 87.87% | 91.70% | 93.60% | 93.12% | 92.57% | ||
| GT concordance | 98.72% | 98.99% | 98.39% | 98.46% | 94.36% | 99.29% | 98.60% | 97.80% | 97.37% | ||
| CLR_L3 | Precision | 92.45% | 93.08% | 88.32% | 90.60% | 83.06% | 92.22% | 93.00% | 92.56% | 94.08% | |
| Recall | 92.54% | 92.15% | 84.35% | 84.28% | 46.33% | 80.10% | 79.30% | 91.30% | 78.33% | ||
| F1 | 92.50% | 92.61% | 86.29% | 87.33% | 59.48% | 85.74% | 85.60% | 91.93% | 85.48% | ||
| GT concordance | 97.24% | 97.68% | 95.65% | 96.17% | 84.27% | 97.76% | 98.80% | 96.78% | 96.00% | ||
| Overall | Precision | 92.08% | 93.28% | 87.01% | 90.37% | 84.76% | 91.90% | 92.73% | 93.28% | 93.17% | |
| Recall | 93.85% | 92.57% | 86.75% | 87.37% | 51.44% | 87.15% | 88.30% | 92.05% | 87.46% | ||
| F1 | 92.95% | 92.92% | 86.83% | 88.83% | 60.29% | 89.38% | 90.30% | 92.66% | 90.07% | ||
| GT concordance | 98.24% | 98.47% | 96.68% | 96.62% | 88.03% | 98.56% | 98.43% | 97.07% | 97.35% | ||
| INS | CLR_L1 | Precision | 89.15% | 87.90% | 72.05% | 80.21% | 69.61% | 86.78% | 92.73% | 93.15% | 88.02% |
| Recall | 92.84% | 90.82% | 91.18% | 89.95% | 20.77% | 86.12% | 71.20% | 79.78% | 86.65% | ||
| F1 | 90.96% | 89.34% | 80.49% | 84.80% | 32.00% | 86.45% | 80.55% | 85.94% | 87.33% | ||
| GT concordance | 96.94% | 97.66% | 91.75% | 92.00% | 68.00% | 98.04% | 61.70% | 87.56% | 96.72% | ||
| CLR_L2 | Precision | 87.98% | 84.74% | 84.15% | 84.62% | 66.76% | 83.15% | 62.85% | 86.10% | 81.14% | |
| Recall | 93.52% | 93.45% | 91.76% | 93.32% | 88.13% | 91.88% | 27.65% | 77.90% | 80.89% | ||
| F1 | 90.67% | 88.88% | 87.79% | 88.75% | 75.97% | 87.30% | 38.40% | 81.80% | 81.02% | ||
| GT concordance | 97.02% | 98.82% | 95.87% | 96.77% | 67.60% | 98.97% | 44.11% | 89.52% | 94.97% | ||
| CLR_L3 | Precision | 89.59% | 87.17% | 78.91% | 81.51% | 60.40% | 88.16% | 91.67% | 87.57% | 89.51% | |
| Recall | 91.44% | 91.48% | 91.71% | 90.40% | 47.28% | 79.83% | 75.48% | 79.93% | 70.42% | ||
| F1 | 90.51% | 89.27% | 84.83% | 85.72% | 53.04% | 83.79% | 82.79% | 83.58% | 78.83% | ||
| GT concordance | 95.07% | 96.21% | 89.24% | 89.86% | 54.27% | 96.23% | 87.68% | 88.65% | 92.63% | ||
| Overall | Precision | 88.91% | 86.60% | 78.37% | 82.11% | 65.59% | 86.03% | 82.42% | 88.94% | 86.22% | |
| Recall | 92.60% | 91.92% | 91.55% | 91.22% | 52.06% | 85.94% | 58.11% | 79.20% | 79.32% | ||
| F1 | 90.71% | 89.16% | 84.37% | 86.42% | 53.67% | 85.85% | 67.25% | 83.77% | 82.39% | ||
| GT concordance | 96.34% | 97.56% | 92.29% | 92.88% | 63.29% | 97.75% | 64.50% | 88.58% | 94.77% |
-
The table presents recall, precision, F1 score (in %), and genotype accuracy (measured by genotype concordance) for all benchmarked SV callers. The highest scores for recall, precision, F1, and genotype accuracy are highlighted in bold. Additionally, the overall average performance for each SV type and library is summarized at the bottom of the table. Benchmarking was conducted using Truvari with the following parameter settings: p = 0.5, P = 0.5, r = 500, and O = 0.01.
