Prediction of Enzyme Mutant Activity Using Computational Mutagenesis and Incremental Transduction

Table 2

Incremental Transductive Learning.

Dataset	Strategy	Avg. Acc.	St.Dev.	Sensitivity	Specificity
HIV-1	T2aNN	75.53	2.70	0.71	0.76
	T2bNN	78.05	2.50	0.75	0.81
	T2aRF	83.46	2.62	0.78	0.82
	T2bRF	86.88	2.55	0.76	0.83
T4	T2aNN	82.69	4.11	0.89	0.50
	T2bNN	82.64	5.14	0.90	0.56
	T2aRF	89.71	3.54	0.93	0.63
	T2bRF	90.97	3.48	0.94	0.67
LAC	T2aNN	76.17	2.88	0.78	0.75
	T2bNN	82.51	2.93	0.80	0.75
	T2aRF	86.54	2.71	0.86	0.80
	T2bRF	90.84	2.87	0.86	0.80

Results of transductive learning algorithms T2a and T2b on HIV-1, T4, and LAC using incremental transductive learning, and selectivity. (the number of folds used for cross-validation is 4 for the HIV-1 dataset and 10 for T4 and LAC.)