Rigorous assessment and integration of the sequence and structure based features to predict hot spots

BMC Bioinformatics

Table 7 Predictive results of different methods when ΔΔG ≥ 2 kcal.mol^-1 is defined as a threshold

Method	Dataset	P	R	F1
Robetta	Training set	0.49	0.44	0.47
	Test set	0.71	0.25	0.37
FOLDEF	Training set	0.45	0.54	0.49
	Test set	0.60	0.26	0.37
Cho's method	Training set	0.58	0.73	0.65
	Test set	0.44	0.65	0.52
Tuncbag's method	Training set	0.64	0.52	0.57
	Test set	0.73	0.59	0.65
Sequence-based SVM	Training set	0.78	0.80	0.79
	Test set	0.65	0.64	0.64

ISSN: 1471-2105