ViralPhos: incorporating a recursively statistical method to predict phosphorylation sites on virus proteins

BMC Bioinformatics

Table 2 Five-fold cross validation results on pSer MDDLogo-clustered SVM models.

SVM model	Number of positive data	Number of negative data	Cost value	Gamma value	Sn	Sp	Acc	MCC
All data	233	233	0.5	0.125	0.76	0.72	0.74	0.48
Subgroup S1	66	66	2	0.125	0.98	0.87	0.93	0.86
Subgroup S2	54	54	8	0.03125	0.94	0.92	0.93	0.87
Subgroup S3	34	34	0.5	0.03125	0.91	0.79	0.85	0.71
Subgroup S4	20	20	2	0.125	0.90	0.80	0.85	0.70
Subgroup S5	15	15	2	0.125	0.87	0.80	0.83	0.66
Subgroup S6	44	44	0.5	0.03125	0.75	0.61	0.68	0.37
Combined performance					0.90	0.82	0.86	0.72

ISSN: 1471-2105