Table 1.
Inclusion of obligate terms has greatest impact in low identity regime
| Identity (%) | Targets | Threshold | Accuracy | Completeness | ||||
|---|---|---|---|---|---|---|---|---|
| Better | Unchanged | Worse | Better | Unchanged | Worse | |||
| 0–5 | 22 | 0.05 | 3 | 18 | 1 | 4 | 17 | 1 |
| 0.10 | 2 | 19 | 1 | 3 | 18 | 1 | ||
| 0.20 | 1 | 21 | 0 | 1 | 21 | 0 | ||
| 5–10 | 69 | 0.05 | 11 | 49 | 9 | 10 | 45 | 14 |
| 0.10 | 7 | 57 | 5 | 5 | 58 | 6 | ||
| 0.20 | 3 | 65 | 1 | 3 | 64 | 2 | ||
| 10–15 | 58 | 0.05 | 11 | 38 | 9 | 10 | 38 | 10 |
| 0.10 | 5 | 47 | 6 | 4 | 46 | 8 | ||
| 0.20 | 1 | 53 | 4 | 1 | 55 | 2 | ||
| 15–20 | 47 | 0.05 | 5 | 38 | 4 | 4 | 40 | 3 |
| 0.10 | 3 | 41 | 3 | 2 | 43 | 2 | ||
| 0.20 | 2 | 43 | 2 | 1 | 44 | 2 | ||
| 20–25 | 35 | 0.05 | 1 | 32 | 2 | 1 | 32 | 2 |
| 0.10 | 1 | 33 | 1 | 1 | 34 | 0 | ||
| 0.20 | 0 | 35 | 0 | 0 | 35 | 0 | ||
| 25–30 | 27 | 0.05 | 3 | 23 | 1 | 2 | 25 | 0 |
| 0.10 | 0 | 27 | 0 | 0 | 27 | 0 | ||
| 30–35 | 20 | 0.05 | 0 | 20 | 0 | 0 | 20 | 0 |
| 35–40 | 15 | 0.05 | 0 | 15 | 0 | 0 | 15 | 0 |
Evaluation of accuracy and completeness for K*Sync default alignments accomplished both with and without obligateness terms on SCOP1.38 test set. Targets are grouped into 5% identity bins, as in Figure 2. The number of ‘targets’ in each bin is reported. Also shown is the number of targets that have alignments that change in quality as a result of inclusion of the obligate terms with default weights. Change is defined as having an ‘accuracy’ or ‘completeness’ deviating by more than a given threshold than that of the default alignment without obligate terms, and are categorized as ‘better’, ‘unchanged’ or ‘worse’.