Table 2.
Evaluation of the structural features in each group of Dataset 2
| Spearman’s ρ | Optimal region | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Feature | HW | HM | HS | LW | LM | LS | HW | HM | HS | LW | LM | LS | Ave. len.a |
| accT | 0.693 | 0.641 | 0.552 | 0.394 | 0.635 | 0.658 | +5:+64 | +2:+20 | −12:+20 | −25:+36 | −22:+72 | −25:+21 | 52.7 ± 26.4 |
| accC | 0.787 | 0.738 | 0.611 | 0.440 | 0.746 | 0.753 | −13:+19 | −18:+18 | −18:+15 | −25:+29 | −22:+20 | −25:+18 | 41.0 ± 8.2 |
| mfeT | 0.728 | 0.642 | 0.536 | 0.420 | 0.654 | 0.648 | −22:+56 | −20:+56 | −18:+56 | −24:+56 | −22:+59 | −22:+56 | 78.8 ± 2.6 |
| mfeC | 0.632 | 0.573 | 0.492 | 0.353 | 0.558 | 0.579 | −23:+66 | −18:+65 | −16:+66 | −25:+67 | −23:+65 | −26:+66 | 88.7 ± 4.3 |
| ensT | 0.747 | 0.658 | 0.550 | 0.428 | 0.668 | 0.663 | −21:+56 | −17:+56 | −17:+56 | −25:+56 | −23:+56 | −22:+56 | 77.8 ± 3.3 |
| ensC | 0.754 | 0.665 | 0.560 | 0.445 | 0.681 | 0.681 | −23:+65 | −21:+65 | −25:+57 | −25:+67 | −22:+65 | −26:+66 | 88.8 ± 3.8 |
| accCopt1 b | 0.767 | 0.736 | 0.610 | 0.404 | 0.736 | 0.746 | – | – | – | – | – | – | – |
| RFopt1 c | 0.703 | 0.645 | 0.539 | 0.372 | 0.621 | 0.610 | – | – | – | – | – | – | – |
| RBSDesigner | 0.597 | 0.561 | 0.441 | 0.300 | 0.539 | 0.534 | – | – | – | – | – | – | – |
| RBSCalculator | 0.631 | 0.569 | 0.476 | 0.300 | 0.595 | 0.594 | – | – | – | – | – | – | – |
HW, HM, HS, LW, LM, and LS are the group codes. The first and second letters of the code indicate the type of promoter and 5′-UTR, respectively (H: high, L: low promoter; S: strong, M: middle, W: weak UTR). The highest ρ value in each group is written in bold.
aThe average length of the optimal region (± standard deviation)
baccC calculated at the optimal region in Dataset 1, that is, −19:+15
cRandom forest regression model trained using Dataset 1