ACNNT3: Attention-CNN Framework for Prediction of Sequence-Based Bacterial Type III Secreted Effectors

Figure 1

ACNNT3 architecture for T3SE prediction. Firstly, 64 1D convolution kernels with a length of 6 are convoluted to generate a feature map, and then a feature map is obtained through a maximum pooling layer. The feature map is then input to the attention and full connection layers, and the two output results are combined to get 66 nodes. Finally, the 66 nodes are fully connected to the two output nodes, and the sigmoid function is used to activate to get the prediction probability of T3SE and non-T3SE.