Fig. 8From: SeqPredNN: a neural network that generates protein sequences that fold into specified tertiary structuresFolded structures of predicted sequences are close to the native structure. The superposition of the control structure predicted from the real sequence (blue) and the crystal structure (tan) is shown, as well as the superposition of the structures predicted for the sequences generated with SeqPredNN, proDCoNN and ProteinSolver (magenta) and the control structure (tan). Structure predictions were performed with AlphaFold (top row) and RoseTTAFold (bottom row). The Ca–Ca RMSD values are shownBack to article page