Skip to main content
. 2019 Mar 4;10:143. doi: 10.3389/fgene.2019.00143

Figure 2.

Figure 2

The schematic diagram of DMfold Architecture, which contains two parts: PU and CU. PU is a deep learning model, mainly responsible for predicting the input RNA sequences as dot-bracket sequences. CU is mainly to correct the prediction dot-bracket sequences and output the prediction secondary structure.