Fig. 1.

Model structure. It includes feature representation based on dna2vec method, two convolutional layers, two pooling layers, bidirectional long short-term memory network layer, attention layer and finally two fully connected layers