Table 2.

The performance of sAMPpred-GAT and the other existing predictors on the independent test XUAMP dataset in terms of ACC, MCC, Sn and Sp^a

Method	ACC	MCC	Sn	Sp	Source
amPEPpy	0.679	0.431	0.400	0.958	Xu et al. (2021)
AMPfun	0.674	0.414	0.406	0.943	Xu et al. (2021)
AMPEP	0.661	0.429	0.330	0.992	Xu et al. (2021)
ADAM-HMM	0.684	0.390	0.521	0.847	Xu et al. (2021)
Ampir	0.563	0.156	0.266	0.859	Xu et al. (2021)
AMPScannerV2	0.568	0.137	0.523	0.613	Xu et al. (2021)
AmpGram	0.564	0.131	0.445	0.682	Xu et al. (2021)
Deep-AMPEP30	0.533	0.183	0.065	1.0	Xu et al. (2021)
CAMP-ANN	0.584	0.182	0.385	0.782	Xu et al. (2021)
sAMPpred-GATb	0.715 ± 0.01	0.464 ± 0.011	0.530 ± 0.038	0.9 ± 0.02	This study

^aThe ACC, MCC, Sn and Sp values of nine compared methods are calculated based on the detailed prediction results of these predictors reported in Xu et al. (2021), which can be downloaded from http://bliulab.net/sAMPpred-GAT/data/.

^bThe reported results of the proposed method are the average and standard deviation after performing the randomness initialization parameters 10 times.

Note: The best performance of each metric is highlighted in bold.