A machine learning strategy for predicting localization of post-translational modification sites in protein-protein interacting regions

Table 5 Resulting performance measures of all six classifiers for all three PTM-specific datasets, using the optimized lists of indices and features as an input

Classifier	Dataset	Relief-F				Information gain
Classifier	Dataset	ACC	AUC	MCC	CPS	ACC	AUC	MCC	CPS
SVM	Acetylation	0.88	0.92	0.78	2.58	0.88	0.90	0.78	2.56
	Phosphorylation	0.91	0.93	0.83	2.67	0.91	0.93	0.84	2.68
	Ubiquitylation	0.88	0.91	0.77	2.56	0.88	0.91	0.77	2.56
		summation			7.81	summation			7.80
k-NN	Acetylation	0.87	0.91	0.74	2.52	0.87	0.91	0.75	2.53
	Phosphorylation	0.89	0.93	0.80	2.62	0.92	0.93	0.84	2.69
	Ubiquitylation	0.80	0.86	0.61	2.27	0.81	0.89	0.65	2.35
		summation			7.41	summation			7.57
RF	Acetylation	1.00	1.00	1.00	3.00	1.00	1.00	1.00	3.00
	Phosphorylation	0.91	0.93	0.82	2.66	0.90	0.93	0.80	2.63
	Ubiquitylation	1.00	1.00	1.00	3.00	1.00	1.00	1.00	3.00
		summation			8.66	summation			8.63
C4.5	Acetylation	0.85	0.87	0.70	2.42	0.88	0.89	0.76	2.53
	Phosphorylation	0.89	0.90	0.78	2.57	0.92	0.93	0.85	2.70
	Ubiquitylation	0.80	0.82	0.61	2.23	0.81	0.82	0.62	2.25
		summation			7.22	summation			7.48
KStar	Acetylation	0.83	0.88	0.65	2.36	0.83	0.89	0.66	2.38
	Phosphorylation	0.87	0.92	0.74	2.53	0.89	0.93	0.79	2.61
	Ubiquitylation	0.71	0.76	0.43	1.90	0.79	0.82	0.57	2.18
		summation			6.79	summation			7.17
MLP	Acetylation	0.85	0.91	0.70	2.46	0.84	0.90	0.67	2.41
	Phosphorylation	0.88	0.92	0.76	2.56	0.91	0.93	0.83	2.67
	Ubiquitylation	0.84	0.90	0.68	2.42	0.83	0.89	0.66	2.38
		summation			7.44	summation			7.46

ISSN: 1471-2105