Prediction of amyloid fibril-forming segments based on a support vector machine

BMC Bioinformatics

Table 2 Prediction performance on the Hexpepset dataset with different properties. All of the results were obtained by LIBSVM with RBF kernel.

Number of properties	Specificity(%)	Sensitivity(%)	Q2^a (%)	MCC^b
1^c	61	70	66	0.31
41^d	80	82	81	0.63
155^e	71	73	72	0.45
531^f	98	15	57	0.24

^aoverall prediction accuracy; ^bMCC, Matthews correlation coefficient; ^cthe property of VINM940104; ^dproperties selected by Genetic Algorithm; ^eproperties that obtained the Q2 higher or equal than 60%; ^fall physicochemical properties in the AAindex.

ISSN: 1471-2105