Fig. 2. Accurate modeling results in identification of motif residues and correlates with high pLDDT values.
a Heatmaps of per residue RMSD (shades of green) for peptide–protein interactions with known motifs (motif set), with the top bar highlighting the motif region in dark blue, followed by corresponding pLDDT heatmaps (shades of brown). Each row corresponds to the network parameters used to generate the model (1 through 5). b Heatmaps for the non-motif set. Note 2b9h, for which reinspection of the literature reveals a motif. c Scatterplots of pLDDT vs. RMSD values for the different datasets, representing peptide per residue values of the AF2 models. For the LNR set, clusters of points are represented by hexagonal bins, colored in shades of yellow according to the number of residues in the cluster. The number of observations in each subsection is noted therein. The motif set is shown in shades of blue and the non-motif set in shades of red. Models with linker are represented by dark colored “x” and models with no linker (separate chains) are represented with pale “+”. Source data are provided as a Source Data file.