Effect of modifying parameter c on PHI's output
| Haplotype | Edit distance | No. of recombinations | Haplotype length (Mbp) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Estimated sequence by PHI | ||||||||||
| c = 10 | c = 102 | c = 103 | c = 10 | c = 102 | c = 103 | c = 10 | c = 102 | c = 103 | Ground truth | |
| APD | 144,835 | 1810 | 9994 | 36 | 10 | 4 | 5.07 | 4.93 | 4.92 | 4.93 |
| DBB | 67,290 | 1377 | 2111 | 13 | 4 | 2 | 5.11 | 5.05 | 5.05 | 5.05 |
| MANN | 131,564 | 35,940 | 38,872 | 37 | 14 | 5 | 5.15 | 5.04 | 5.04 | 5.03 |
| QBL | 125,655 | 3343 | 13,034 | 46 | 17 | 4 | 5.03 | 4.90 | 4.90 | 4.90 |
| SSTO | 125,770 | 4637 | 13,600 | 68 | 24 | 5 | 5.17 | 5.04 | 5.05 | 5.05 |
[i] The table presents output statistics obtained by running PHI using three different choices of the recombination penalty parameter c: 10, 100, and 1000. We calculated (1) accuracy, that is, edit distance values between the estimated haplotype sequences and the ground-truth assemblies; (2) count of recombinations in the estimated haplotype sequences, and (3) lengths of the estimated sequences. The last column shows the length of the ground-truth sequences.