Table 3.

Independent study of isoform-level circular RNA prediction for mouse circRNAs based on the models trained on human circRNAs

Method	Acc	Prec	Sens	Spec	F1	MCC	AUC
SVM	0.7328	0.7742	0.8108	0.6011	0.7921	0.4196	0.7059
RF	0.7186	0.7393	0.8523	0.4929	0.7918	0.3733	0.6726
Att-CNN	0.7264	0.7452	0.7957	0.6330	0.7696	0.4352	0.7143
Att-RNN	0.7030	0.7189	0.7930	0.5816	0.7541	0.3844	0.6873
PredicircRNA	0.5696	0.6218	0.5056	0.6437	0.5577	0.1501	0.6067
nRC	0.7410	0.7662	0.8455	0.5647	0.8039	0.4298	0.8097
circDeep	0.6140	0.7495	0.6982	0.7509	0.7229	0.4491	0.7669
DeepCirCode	0.8129	0.9271	0.7620	0.8989	0.8365	0.6392	0.8304
JEDI	0.8654	0.9074	0.8749	0.8493	0.8909	0.7162	0.8621