Table 4.
Comparison of SWER, MEC and runtimes for SDhaP and CPLEX on simulated diploid data with 1% error rate
Data | Simulated data | ||||||||
---|---|---|---|---|---|---|---|---|---|
SDhaP | CPLEX(optimal) | CPLEX(heuristic) | |||||||
MEC | SWER | time | MEC | SWER | time | MEC | SWER | time | |
l 103, c 10 | 100 | 0.001 | 1 | 100 | 0.001 | 192 | 100 | 0.001 | 57 |
l 103, c 20 | 215 | 0.001 | 4 | 215 | 0.001 | 1320 | 215 | 0.001 | 373 |
l 103, c 30 | 291 | 0.001 | 6 | 291 | 0.001 | 1241 | 291 | 0.001 | 910 |
l 104, c 10 | 978 | 0.008 | 14 | - | - | - | 972 | 0.008 | 4505 |
l 104, c 20 | 2039 | 0.004 | 33 | - | - | - | 2039 | 0.004 | 89811 |
l 104, c 30 | 2988 | 0.004 | 68 | - | - | - | - | - | - |
l 105, c 10 | 10356 | 0.008 | 324 | - | - | - | - | - | - |
l 105, c 20 | 19975 | 0.007 | 713 | - | - | - | - | - | - |
l 105, c 30 | 29967 | 0.005 | 1810 | - | - | - | - | - | - |
MEC, SWER and running times (in seconds) for SDhaP, RefHap, HAPCUT and HapTree algorithms for simulated data of different lengths (l) and with different coverages (c). The data contains a fixed 1% fraction of genotyping errors. SDhaP is more accurate in terms of MEC and SWER and faster by almost an order of magnitude compared to other schemes for longer blocks.