Prediction of the functional class of metal-binding proteins from sequence derived physicochemical properties by support vector machine approach

Table 1 Statistics of the datasets and prediction accuracy of individual class of metal-binding proteins and that of all metal-binding proteins. The predicted results are given in TP (true positive), FN (false negative), TN (true negative), FP (false positive), sensitivity SE = TP/(TP + FN) (accuracy for class members), specificity SP = TN/(TN + FP) (accuracy for non-members), and overall accuracy Q = (TN + TP)/(TP + FN + TN + FP). The number of members and non-members in the testing and independent evaluation sets is TP + FN or TN + FP respectively

Metal-Binding Protein Classes	Training set		Testing set				Independent evaluation set
	positive	negative	positive		negative		positive			negative			Q(%)
			TP	FN	TN	FP	TP	FN	SE(%)	TN	FP	SP(%)
Calcium-binding	1816	4389	2055	245	8039	1906	1130	180	86.3%	6373	851	88.2%	87.9%
Cobalt-binding	568	2151	456	2	13407	10	360	81	81.6%	7809	7	99.9%	98.9%
Copper-binding	652	1999	417	109	13115	270	390	77	83.5%	7587	146	98.1%	97.3%
Iron-binding	3128	3428	4869	290	3992	675	1104	71	94.0%	6328	598	91.4%	91.7%
Magnesium-binding	2583	4023	2307	594	7267	115	3412	792	81.2%	5848	805	87.9%	85.3%
Manganese-binding	1608	3099	1146	217	10841	1086	1061	182	85.4%	7148	415	94.5%	93.2%
Nickel-binding	407	2001	95	2	13576	138	156	45	77.6%	7816	65	99.2%	98.6%
Potassium-binding	408	1845	489	10	13789	8	301	32	90.4%	7847	4	99.9%	99.6%
Sodium-binding	777	2010	338	1	13591	30	410	41	90.9%	7831	11	99.9%	99.4%
Zinc-binding	2731	6416	6610	569	5931	360	4616	1546	74.9%	6289	127	98.0%	86.7%
All metal-binding	5013	3101	11806	1015	4217	522	12070	3391	78.1%	4529	617	88.0%	80.6%

ISSN: 1471-2105