PredPhos: an ensemble framework for structure-based prediction of phosphorylation sites

Table 1 Performance of the two-step feature selection method

	AUC	Accuracy	Recall	Specificity	CC	F1
CK2
Optimal features	0.877	0.963	0.433	0.992	0.433	0.429
All features	0.842	0.954	0.350	0.986	0.370	0.366
MAPK
Optimal features	0.839	0.952	0.483	0.973	0.480	0.480
All features	0.833	0.959	0.400	0.985	0.424	0.423
PKA
Optimal features	0.858	0.948	0.375	0.980	0.426	0.432
All features	0.846	0.926	0.335	0.959	0.279	0.310
PKC
Optimal features	0.857	0.952	0.303	0.985	0.396	0.363
All features	0.821	0.948	0.226	0.984	0.282	0.274
SRC
Optimal features	0.900	0.951	0.558	0.973	0.510	0.499
All features	0.890	0.946	0.241	0.985	0.317	0.294

ISSN: 2241-5793