Table 2.
Topological ancestor–descendant and sibling relationship reconstruction on simulated data
| Samples | Cov | Trees | % SSNVs | % AD | % AD-Ord | % AD-Corr | % AD →Sib (−priv) | % Sib | %Sib-Corr | % Sib →AD (−priv) |
|---|---|---|---|---|---|---|---|---|---|---|
| 5 | t-VAF | 94 | 98.9 | 99 | 41.5 | 100 | 30.2 (2.6) | 98.2 | 83.2 | 7.3 (1.5) |
| 98 | 99.5 | 99.5 | 40.4 | 100 | 26.3 | 99.2 | 84.8 | 6.6 (0.9) | ||
| 10K | 95 | 98.8 | 98.7 | 40.4 | 99.9 | 30.5 (2.5) | 97.8 | 82.5 | 7.7 (1.4) | |
| 98 | 99.2 | 98.7 | 39 | 100 | 27.4 | 98.8 | 84.7 | 7 (1.3) | ||
| 1K | 95 | 98.6 | 97.7 | 39.7 | 100.0 | 30.5 (2.3) | 97.1 | 82.8 | 7.2 (1.3) | |
| 98 | 99.4 | 99.5 | 38.1 | 99.7 | 29.2 | 98.9 | 83.9 | 7.5 (2.2) | ||
| 100 | 97 | 96.8 | 93.2 | 38.7 | 99.8 | 31.4 (3.5) | 93.1 | 81.8 | 7.9 (1.7) | |
| 97 | 97.3 | 94.8 | 35.2 | 99.7 | 30.1 | 95.3 | 83.7 | 7.2 (2.2) | ||
| 10 | t-VAF | 97 | 97.5 | 96.2 | 58.8 | 100 | 43.1 (2.2) | 96.9 | 94.4 | 3.3 (0.2) |
| 95 | 97.7 | 96.5 | 58.2 | 100 | 38.1 (1.9) | 96.7 | 93.5 | 4.2 (1.4) | ||
| 10K | 98 | 96.5 | 94.4 | 57.3 | 100 | 44.9 (4) | 95.2 | 94 | 3.6 (0.4) | |
| 96 | 97 | 96.1 | 56.5 | 100 | 39.8 (2.3) | 96.2 | 93.3 | 4.3 (1.2) | ||
| 1K | 98 | 97.4 | 96.4 | 57.5 | 99.9 | 45.2 (3.9) | 96.8 | 94.1 | 3.7 (0.4) | |
| 96 | 97 | 95.7 | 57.4 | 99.9 | 38.7 (2.4) | 96 | 92.7 | 4.8 (1.5) | ||
| 100 | 93 | 90.8 | 75.7 | 52.3 | 99.7 | 38.6 (4.4) | 78.0 | 93.9 | 3.4 (0.7) | |
| 91 | 90.5 | 80.4 | 47.9 | 99.5 | 40.3 (5.7) | 83.3 | 92.8 | 4.4 (1.5) | ||
| 15 | t-VAF | 99 | 91.7 | 85 | 63.1 | 100 | 40.8 (2.2) | 88 | 96.6 | 2 (0.2) |
| 96 | 92.1 | 87.5 | 61.7 | 100 | 41.5 (2) | 89.1 | 96.9 | 2 (0.3) | ||
| 10K | 98 | 92.3 | 85.3 | 61.2 | 100 | 44.3 (3.5) | 87.6 | 96.2 | 2.4 (0.3) | |
| 100 | 90.6 | 83.6 | 59.1 | 100 | 41.8 (2.6) | 85.8 | 96.6 | 2.1 (0.3) | ||
| 1K | 93 | 94.7 | 83.5 | 61.5 | 99.9 | 42.7 (4.3) | 85.4 | 96.1 | 2.6 (0.6) | |
| 100 | 91.8 | 85.4 | 59.1 | 100 | 43.5 (2.9) | 87.8 | 96.7 | 2.1 (0.3) | ||
| 100 | 92 | 81.1 | 55.2 | 48.4 | 99.4 | 37.4 (5) | 61.3 | 96.1 | 2.1 (0.3) | |
| 98 | 81.8 | 57.1 | 46.1 | 99.8 | 36.5 (2.2) | 63.1 | 96.2 | 2.2 (0.4) |
Results are shown for 5, 10, and 15 samples given true VAFs, 10,000 ×, 1,000 ×, and 100 × coverage data (without CNVs) obtained with localized (top row in each pair) and randomized (bottom row) sampling. All values are averaged over the number of reconstructed trees (Trees) out of 100 simulated trees. The following metrics are presented: SSNVs present in the tree (% SSNVs), ancestor–descendant pairs of mutations in the tree (AD), ordered AD pairs (AD-Ord), correctly ordered AD pairs (AD-Corr), unordered AD pairs reconstructed as siblings (AD →Sib) (with and without private mutation nodes), sibling pairs of mutations in the tree (Sib), correctly reconstructed sibling pairs (Sib-Corr), sibling pairs reconstructed as AD (Sib →AD) (with and without private mutation nodes). AD, ancestor–descendant; CNV, copy number variant; Corr, correctly ordered; Cov, coverage; Ord, ordered; priv, private mutation nodes; Sib, sibling; SSNV, somatic single nucleotide variant; t-VAF, true variant allele frequency; VAF, variant allele frequency.