Table 1.

AUC_b scores of models (w2v + LSTM + attention) with different input length.

Length (bp)	Am	Cm	Gm	Tm	m1A	m5C	m5U	m6A	m6Am	m7G	Ψ	I	Mean	Median
101	0.7900	0.8224	0.9108	0.8596	0.8300	0.9108	0.9196	0.8496	0.8944	0.6204	0.8228	0.6040	0.8195	0.8398
51	0.7272	0.8452	0.9324	0.8832	0.8068	0.9236	0.9560	0.8336	0.9272	0.6012	0.8420	0.6304	0.8257	0.8436
21	0.7276	0.7984	0.9468	0.8396	0.7624	0.9360	0.9404	0.7848	0.8460	0.6004	0.7800	0.6084	0.7976	0.7916

Please note that our predictive model requires information of neighboring sequences even if they do not form part of the transcript. Bold indicates the best performance in comparison.