Table 1 Accuracy of several decoding methods on simulated data
From: Probabilistic approaches to alignment with tandem repeats
Alignment | Repeat | Block | |||
|---|---|---|---|---|---|
Algorithm | error | sn. | sp. | sn. | sp. |
SFF marginalized | 3.37% | 95.97% | 97.78% | 43.07% | 44.87% |
SFF posterior | 3.53% | 95.86% | 97.87% | 42.70% | 47.37% |
SFF block | 3.51% | 93.09% | 98.07% | 36.50% | 41.67% |
SFF block Viterbi | 3.91% | 93.26% | 97.96% | 35.77% | 40.66% |
SFF Viterbi | 4.04% | 95.29% | 97.85% | 42.70% | 48.95% |
TANTAN block | 5.05% | 61.38% | 97.48% | 0.00% | 0.00% |
TANTAN block Viterbi | 6.17% | 67.86% | 96.51% | 0.00% | 0.00% |
SFF marginalized∗ | 3.02% | 98.93% | 99.64% | 77.01% | 76.17% |
SFF posterior∗ | 3.42% | 98.84% | 99.51% | 75.91% | 80.93% |
SFF block∗∗ | 3.21% | 97.70% | 99.87% | 80.66% | 94.44% |
SFF block Viterbi∗∗ | 3.71% | 98.12% | 99.85% | 81.75% | 92.18% |
SFF Viterbi∗ | 3.94% | 98.54% | 99.45% | 75.55% | 83.47% |
TANTAN block† | 3.42% | 60.45% | 99.90% | 0.36% | 0.46% |
TANTAN block Viterbi† | 3.83% | 61.74% | 99.88% | 0.00% | 0.00% |
Context | 5.98% | ||||
Muscle | 5.62% | ||||
3-state posterior | 4.41% | ||||
3-state posterior with masked repeats†† | 5.03% | 99.23% | 74.16% | 7.66% | 7.24% |
3-state Viterbi (baseline) | 4.78% | ||||
SFF marginalized◇ | 3.63% | 96.03% | 97.74% | 42.70% | 43.33% |
SFF marginalized◇◇ | 3.36% | 95.99% | 97.81% | 40.88% | 43.08% |