Table 9.
SOV’99 scores of Q8 prediction using eCRRNN on 4 datasets: CB513, CASP10, CASP11 and CASP12
| Type | CB513 | CASP10 | CASP11 | CASP12 |
|---|---|---|---|---|
| SOV L | 0.611 | 0.629 | 0.595 | 0.538 |
| SOV H | 0.929 | 0.924 | 0.908 | 0.907 |
| SOV T | 0.599 | 0.67 | 0.605 | 0.531 |
| SOV E | 0.882 | 0.884 | 0.86 | 0.829 |
| SOV S | 0.351 | 0.362 | 0.327 | 0.256 |
| SOV B | 0.078 | 0.11 | 0.077 | 0.027 |
| SOV G | 0.419 | 0.416 | 0.339 | 0.297 |
| SOV I | 0 | 0 | 0 | 0 |
| SOV | 0.725 | 0.747 | 0.722 | 0.684 |
| SOV mean | 0.723 | 0.740 | 0.738 | 0.698 |
SOV L,SOV H, SOV T, SOV E, SOV S, SOV B, SOV G and SOV I represent the prediction SOV score on a constituent secondary structure type L, H, T, E, S,B, G and I respectively. SOV represents the SOV score on different dataset and SOV mean is mean value of the SOV score on sequence level