Skip to main content
. 2020 May 13;5:100035. doi: 10.1016/j.gene.2020.100035

Fig. 4.

Fig. 4

Splice2Deep learning model. It takes DNA sequence as input embedded in 2D (either 4 × L or 64 × L), apply k motif detectors (filters), max pooling, flatten, fully connected layer using SoftMax to output scores.