Prediction and classification of aminoacyl tRNA synthetases using PROSITE domains

BMC Genomics

Table 3 The performance of SVM modules developed for discriminating class-1 and class-2 aaRSs.

S. No.	Name of Approach	Sensitivity	Specificity	Accuracy	MCC
1	Domain-based model	100	53.85	74.64	0.56
2	Amino acid composition	68	64.62	66.09	0.33
3	Dipeptide composition	85.27	92.31	88.99	0.79
4	Hybrid approach¹*	81.09	83.08	82.18	0.65
5	Hybrid approach²**	94.55	92.31	93.19	0.87

* Hybrid approach¹ SVM model based on amino acid composition domain features
** Hybrid approach² SVM model based on selected dipeptide composition and domain features

ISSN: 1471-2164