Automatic prediction of catalytic residues by modeling residue structural neighborhood

BMC Bioinformatics

Table 7 Comparison with the structure-based approaches by Chea et al. [13] and Youn et al. [10] on their benchmark datasets.

Method	Datasets of competing methods
		HA superfamily[13]	EF fold[10]	EF superfamily[10]	EF family[10]
Competing methods	R	29.3	51.1	53.9	57.0
	P	16.5	17.1	16.9	18.5
			SVM_P 24_1D_1-3, 3D_1-6
Equal P	R	63.4	64.2	67.3	61.7
Equal R	P	30.9	22.1	22.5	20.9

For each dataset we report the recall obtained by our predictor at a precision equal to that of the competing method and the precision at equal recall. Results are obtained without including ligand information.

ISSN: 1471-2105