Prediction of nuclear proteins using nuclear translocation signals proposed by probabilistic latent semantic indexing

Table 2 Performance comparison

Training Data Set (by five-fold cross-validation)
Method	tp	tn	fp	fn	Sens	Spec	Acc	MCC
PSLNuc	2,317	2,030	576	525	0.815	0.779	0.797	0.595
NUCLEO	2,157	1,924	682	685	0.759	0.760	0.749	0.497
Independent Test Data Set
	tp	tn	fp	fn	Sens	Spec	Acc	MCC
PSLNuc	452	258	140	112	0.805	0.646	0.739	0.457
NUCLEO	430	246	152	134	0.760	0.620	0.700	0.380
PredictNLS	153	369	29	411	0.270	0.930	0.540	0.250
NucPred	376	233	165	188	0.670	0.590	0.630	0.250

ISSN: 1471-2105