Table 3. Five-fold cross-validation performance for five SVM models trained from MDDLogo-identified motifs.

Dataset	Number of positive data	Number of negative data	Sn	Sp	Acc	MCC	AUC
All data	710	5,676	0.72	0.73	0.73	0.38	0.78
Palm1	112	895	0.82	0.83	0.83	0.50	0.87
Palm2	104	831	0.81	0.81	0.81	0.48	0.86
Palm3	183	1463	0.83	0.84	0.84	0.53	0.89
Palm4	107	856	0.79	0.81	0.81	0.47	0.86
Palm5	204	1631	0.73	0.73	0.73	0.39	0.79
Combined result	710	5676	0.79	0.80	0.80	0.45	0.85