Table 3.
Statistics for Prediction with the Default Parameters
| Statistics | ||||
| CC | OQ | PQ | Z | |
| Training set | ||||
| eve2 | 0.50 | 0.38 | 0.59 | 4.7 |
| hairy7 | 0.39 | 0.31 | 0.48 | 8.6 |
| hairy6 | 0.38 | 0.32 | 0.20 | 9.5 |
| tll | 0.35 | 0.28 | 0.46 | 4.5 |
| enint | 0.33 | 0.28 | 0.37 | 4.5 |
| eve37 | 0.32 | 0.29 | 0.30 | 7.7 |
| kr730 | 0.30 | 0.29 | 0.27 | 6.5 |
| iab2 | 0.24 | 0.16 | 0.47 | 8.5 |
| ftzprox | 0.23 | 0.23 | 0.24 | 3.2 |
| sal | 0.18 | 0.19 | 0.19 | 4.1 |
| Experimental set | ||||
| eve4+6 | 0.61 | 0.52 | 0.64 | 9.3 |
| runt5 | 0.15 | 0.15 | 0.20 | 9 |
-
The default parameters for the group of 10 sequences were 7–9 bp for word length, 0–1 mismatches allowed, and 0.24 for default coverage (see Table 2). The same parameters were applied to the simulated experimental set: eve4+6 and runt5 regions. In comparison to individually trained parameters (Table 2), overall correlation was lower in all cases.
