Fig. 2From: DEEPSEN: a convolutional neural network based method for super-enhancer predictionThe architecture of DEEPSEN-2L. DEEPSEN-2L consists of the input layer, the 1st convolutional layer (including the 1st max-Pooling), the 2nd convolutional layer (including the 2nd max-Pooling), fully connected layer (including dropout), and output softmax layerBack to article page