Predicting Interactions between Virus and Host Proteins Using Repeat Patterns and Composition of Amino Acids

Table 1

Results of 10-fold cross validation of SVM model on 1,072 PPIs between 36 RNA viruses and 812 human proteins with different ratios of positive to negative instances.

P : N	Dataset	Sn (%)	Sp (%)	Acc (%)	PPV (%)	NPV (%)	MCC	AUC
1 : 1	1	88.24	97.34	92.79	97.07	89.22	0.859	0.963
	2	81.03	94.36	87.7	93.49	83.26	0.761	0.931
	3	77.74	94.04	85.89	92.88	80.86	0.728	0.926
	mean  SD	82.34  4.39	95.25  1.49	88.79  2.92	94.48  1.85	84.45  3.51	0.78  0.06	0.94  0.02
1 : 2	1	64.89	97.34	86.52	92.41	84.72	0.693	0.893
	2	58.31	97.57	84.48	92.31	82.4	0.646	0.886
	3	63.64	96.08	85.27	89.04	84.09	0.661	0.891
	mean  SD	62.28  2.85	97  0.65	85.42  0.84	91.25  1.57	83.74  0.98	0.67  0.02	0.89  0.00
1 : 3	1	46.24	98.28	85.27	89.94	84.58	0.58	0.850
	2	46.87	98.59	85.66	91.72	84.77	0.59	0.863
	3	49.37	97.28	85.31	85.83	85.22	0.576	0.858
	mean  SD	47.49  1.35	98.05  0.56	85.41  0.18	89.16  2.47	84.86  0.27	0.58  0.01	0.86  0.01

Sn: sensitivity, Sp: specificity, Acc: accuracy, PPV: positive predictive value, NPV: negative predictive value, MCC: Matthews correlation coefficient, and AUC: the area under the ROC curve.