Table 5 Open in new tab Performance...

Table 5

Performance comparison with the latest accessible methods

Method name	Algorithm	Encoding scheme	Species	Datasize	Sn	Sp	MCC	AUC
SRAMP [14]	RF	one-hot, KNN score spectrum	H. sapiens, mm10	55 706 (full transcript mode) and 46992 (mature mRNA mode)	0.44	0.9	0.29	0.794
WHISTLE [25]	SVM	NCP,GNF, Genome-derived features	H. sapiens	20 516, 17 383	0.62	0.9	0.49	0.83
DeepPromise [6]	CNN	ENAC, one-hot and RNA word embedding	H. sapiens and Mus musculus	44 901, 11 656 and 5233	0.39	0.9	0.252	0.76
DeepOME [60]	CNN+BLSTM	one-hot	H. sapiens	3052	0.97	1.00	0.93	0.99
CapNetwork	CNN+BLSTM+CapsuleNetwork	RNA word embedding	mm10	207 010	0.12	0.95	0.13	0.61
Adaptive-M6A	CNN+BLSTM+attention	RNA word embedding	BDGP6	31 952	0.99	1.00	0.97	1.00
			danRer10	88 779	0.99	0.96	0.95	0.99
			E. coli	5240	0.75	0.97	0.77	0.98
			mm10	207 010	0.97	0.13	0.09	0.79
			TAIR10	10 569	0.73	0.61	0.34	0.71
			sacCer3	5487	0.77	0.73	0.50	0.81

Method name	Algorithm	Encoding scheme	Species	Datasize	Sn	Sp	MCC	AUC
SRAMP [14]	RF	one-hot, KNN score spectrum	H. sapiens, mm10	55 706 (full transcript mode) and 46992 (mature mRNA mode)	0.44	0.9	0.29	0.794
WHISTLE [25]	SVM	NCP,GNF, Genome-derived features	H. sapiens	20 516, 17 383	0.62	0.9	0.49	0.83
DeepPromise [6]	CNN	ENAC, one-hot and RNA word embedding	H. sapiens and Mus musculus	44 901, 11 656 and 5233	0.39	0.9	0.252	0.76
DeepOME [60]	CNN+BLSTM	one-hot	H. sapiens	3052	0.97	1.00	0.93	0.99
CapNetwork	CNN+BLSTM+CapsuleNetwork	RNA word embedding	mm10	207 010	0.12	0.95	0.13	0.61
Adaptive-M6A	CNN+BLSTM+attention	RNA word embedding	BDGP6	31 952	0.99	1.00	0.97	1.00
			danRer10	88 779	0.99	0.96	0.95	0.99
			E. coli	5240	0.75	0.97	0.77	0.98
			mm10	207 010	0.97	0.13	0.09	0.79
			TAIR10	10 569	0.73	0.61	0.34	0.71
			sacCer3	5487	0.77	0.73	0.50	0.81