VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines

BMC Bioinformatics

Table 1 VaxiJen models validation.

model	validation	AUC_ROC^a	threshold ^b	accuracy%^c	sensitivity%^d	specificity%^e
	LOO-CV	0.883	0.5	80	79	81
bacterial	test set	0.726	0.5	70	76	64
	LOO-CV (mean)^g	0.899	0.5	83	81	85
	LOO-CV	0.937	0.5	87	91	82
viral	test set	0.743	0.4	70	84	56
	LOO-CV (mean)^g	0.810	0.5	73	74	71
	LOO-CV	0.964	0.5	89	94	84
tumour	test set	0.930	0.5	86	96	76
	LOO-CV (mean)^g	0.911	0.5	82	78	86

^aThe area under the curve (AUC_ROC) is a quantitative measure of the predictive ability and varies from 0.5 for a random prediction to 1.0 for a perfect prediction.
^bThe threshold of the highest accuracy.
^cAccuracy = (true antigens + true non-antigens)/total.
^dSensitivity = true antigens/all antigens.
^eSpecificity = true non-antigens/all non-antigens.
^gMean values of five training sets.

ISSN: 1471-2105