Table 4.

Prediction results with and without Seq-GAN on the independent test dataset.

Layer	Method	$\mathbf{Acc}$	$\mathbf{Sn}$	$\mathbf{Sp}$	$\mathbf{MCC}$
First Layer (Enhancer Identification)	Without Seq-GAN	0.772	0.799	0.746	0.578
	With Seq-GAN	0.784	0.811	0.758	0.567
Second Layer (Enhancer Strength Identification)	Without Seq-GAN	0.724	0.917	0.531	0.537
	With Seq-GAN	0.749	0.961	0.537	0.505