Table 2.

Comprehensive comparisons with three existing enhancer prediction methods.

Comparison Targets	Method		ACC	AUC	Sens	Spec	GM
EnhancerFinder [40]	EnhancerFinder		\	0.960	\	\	\
	HACF		0.915	0.962	0.904	0.928	0.916
DEEP [29]	Heart	DEEP	0.822	\	0.802	0.824	0.812
		HACF	0.941	0.919	0.964	0.690	0.734
	Liver	DEEP	0.745	\	0.740	0.755	0.741
		HACF	0.923	0.918	0.935	0.664	0.691
	Brain	DEEP	0.853	\	0.832	0.855	0.843
		HACF	0.940	0.918	0.960	0.698	0.744
iEnhancer-2L [32]	Layer I	iEnhancer-2L	0.769	0.850	0.781	0.759	\
		HACF	0.837	0.907	0.846	0.828	0.838
	Layer II	iEnhancer-2L	0.619	0.660	0.622	0.618	\
		HACF	0.714	0.789	0.713	0.716	0.715

HACF: hybrid abelian complexity features; ‘\’ represents “not provided by original publications”.