Fig. 6: Visualization of the features that DiffNets find important for increased stability of M182T and M182S variants. | Nature Communications

Fig. 6: Visualization of the features that DiffNets find important for increased stability of M182T and M182S variants.

From: Deep learning the structural determinants of protein biochemical properties by comparing structural ensembles with DiffNets

Fig. 6

a Crystal structure of TEM β-lactamase (PDB ID: 1JWP) overlaid with dotted lines that indicate distances between two atoms that change in a way that is strongly correlated with an increased DiffNet output label. Red indicates the atoms move closer together as the output label increases, blue indicates atoms moving away from each other. The mutated residue is highlighted with a yellow sphere. Protein atoms are colored cyan if they are near the mutation, which indicates that they were included in the classification task and considered for the distance correlation calculation. b Rotated inset of a showing DiffNet predicted packing at the interface of helix 9 and the adjacent β-sheet. Residues with chemical-shift perturbations in M182S relative to wild-type are shown in deep olive.

Back to article page