Table 4 |
|||
|
Performance comparison of promoter prediction methods on the human genome |
|||
|
Methods |
Specificity |
Sensitivity |
F-measure |
|
|
|||
|
FirstEF |
0.415 |
0.448 |
0.431 |
|
DragonGSF |
0.686 |
0.357 |
0.470 |
|
McPromoter |
0.623 |
0.204 |
0.307 |
|
EP3 |
0.565 |
0.413 |
0.477 |
|
Profisi |
0.604 |
0.392 |
0.475 |
|
ARTS |
0.672 |
0.381 |
0.486 |
|
Eponine |
0.671 |
0.367 |
0.475 |
|
ProSom |
0.573 |
0.414 |
0.481 |
|
PNNP |
0.593 |
0.433 |
0.501 |
|
FSPP(ReliefF-300) |
0.657 |
0.528 |
0.585 |
|
FSPP(Wrapper-SVM) |
0.662 |
0.536 |
0.592 |
|
|
|||
|
These methods include nine existing promoter prediction methods and our prediction methods based on the filter and wrapper feature selection. |
|||
|
Gan et al. BMC Bioinformatics 2012 13:4 doi:10.1186/1471-2105-13-4 |
|||
