Skip to main content
. 2020 Nov 30;9(12):854. doi: 10.3390/antibiotics9120854

Figure 3.

Figure 3

Recurrent neural networks used for peptide property prediction. A representation of each of the amino acids is input to one unit of the recurrent network. The last unit’s output contains information considering the output of the previous ones, therefore considering the whole sequence. On the left a long-short term memory networks (LSTM) unit representation and on the right a gated recurrent units (GRU) unit representation. (Created with BioRender).