RF-Phos: A Novel General Phosphorylation Site Prediction Tool Based on Random Forest

Table 6

Scoring metrics using 10-fold cross validation.

Methods	Residue = S
	AUC	Sen (%)	Sp (%)	MCC
NetPhosK	0.63	50.9	67.8	0.08
GPS 2.1	0.74	33.1	93.3	0.20
Swaminathan	0.70	31.3	88.7	0.13
NetPhos	0.70	34.1	86.7	0.12
PPRED	0.75	32.3	91.6	0.17
Musite	0.81	41.4	93.7	0.25
PhosphoSVM	0.84	44.4	94.0	0.30
RF-Phos	0.88	84.0	85.0	0.65
Methods	Residue = T
	AUC	Sen (%)	Sp (%)	MCC
NetPhosK	0.60	62.0	56.8	0.07
GPS 2.1	0.70	38.1	92.3	0.20
Swaminathan	0.72	28.0	92.5	0.14
NetPhos	0.66	34.3	83.7	0.09
PPRED	0.73	30.3	91.0	0.13
Musite	0.78	33.8	94.8	0.22
PhosphoSVM	0.82	37.3	95.0	0.25
RF-Phos	0.90	83.0	94.0	0.70
Methods	Residue = Y
	AUC	Sen (%)	Sp (%)	MCC
NetPhosK	0.60	39.5	74.2	0.08
GPS 2.1	0.61	34.5	78.9	0.08
Swaminathan	0.62	60.5	57.0	0.09
NetPhos	0.65	34.7	84.5	0.13
PPRED	0.70	43.0	82.7	0.17
Musite	0.72	38.4	86.7	0.18
PhosphoSVM	0.74	41.9	87.3	0.21
RF-Phos	0.91	83.0	88.0	0.70