Table 1. An evaluation of performances of some of the well-known models for glycosylation prediction on prokaryotic glycoproteins.

Type of Glycosylation	Prediction Tools	Threshold	Sensitivity (%)	Specificity (%)	Accuracy (%)	MCC (%)
N-linked	NetNglyc1	0.5	81.75	10.16	34.41	−0.11
		0.6	50.79	42.68	45.43	−0.06
		0.7	15.87	76.02	55.65	−0.09
		0.9	0.79	98.37	65.32	−0.03
	EnsembleGly3	0.3	92.86	0.41	31.72	−0.2
		0.5	90.48	1.63	31.72	−0.18
		0.7	79.37	10.98	34.14	−0.13
		0.9	51.59	47.15	48.66	−0.01
O-linked	NetOglyc2	0	8.38	95.64	87.96	0.05
	EnsembleGly3	0	9.5	93.37	86	0.03

Footnotes: (1: http://www.cbs.dtu.dk/services/NetNGlyc/, 2: http://www.cbs.dtu.dk/services/NetOGlyc-3.0/, 3: http://turing.cs.iastate.edu/EnsembleGly/).