Identification of properties important to protein aggregation using feature selection

BMC Bioinformatics

Table 5 Results of testing ProA-SVM and ProA-RF on the SP dataset

Predictor	Dataset	TP	FN	FP	TN	Ac	MCC
ProA-SVM	EUKSIG.reduc	1022	103	189	936	0.8702	0.7426
	EUKANC.reduc	55	12	10	56	0.8346	0.6695
	GRAM+SIG.reduc	118	51	33	136	0.7515	0.5058
	GRAM-SIG.reduc	286	64	73	277	0.8043	0.6088
	Total	1481	230	305	1405	0.8436	0.6879
ProA-RF	EUKSIG.reduc	896	229	297	828	0.7662	0.5334
	EUKANC.reduc	60	7	9	57	0.8797	0.7597
	GRAM+SIG.reduc	127	42	38	131	0.7633	0.5268
	GRAM-SIG.reduc	290	60	104	246	0.7657	0.5357
	Total	1373	338	448	1262	0.7702	0.5416

TP: the number of True Positive samples; FN: the number of False Negative samples; FP: the number of false positive samples and TP: the number of true positive samples. Ac, Accuracy; MCC, Matthews correlation coefficient.

ISSN: 1471-2105