Prediction of nuclear proteins using SVM and HMM models

BMC Bioinformatics

Table 1 The performance of SVM models using various types of composition.

Composition Type	Sensitivity	Specificity	Accuracy	MCC
Amino Acids	81.33	81.75	81.64	0.59
Dipeptides	82.03	83.11	82.83	0.61
SAAC 2-parts (equal)	82.18	82.68	82.55	0.61
NT15+R	83.03	85.76	85.05	0.65
NT25+R	85.50	85.46	85.47	0.66
NT35+R	83.69	83.89	83.84	0.63
CT15+R	77.41	82.06	80.84	0.56
CT25+R	80.29	80.93	80.77	0.57
CT35+R	80.19	81.22	80.95	0.57
SAAC 3-parts (equal)	81.48	85.04	84.11	0.63
NT15+R+CT15	83.62	83.65	83.64	0.63
NT25+R+CT25	*83.69*	*85.91*	*85.33*	*0.66*
NT35+R+CT35	83.77	84.03	83.96	0.63
NT45 +R+CT45	82.77	83.95	83.64	0.62
SAAC 4-parts (equal)	83.80	83.47	83.55	0.63

In split amino acid composition whole protein was divided into X {X = 2,3,4} equal parts; amino acid composition of each fragments was determined individually and concatenated together to make final input vector of dimension 20*X. NT15 = amino acid composition of N-terminal 15 residues, CT15 = amino acid composition of C-terminal 15 residues and so on; R = amino acid composition of remaining residues.

ISSN: 1471-2105