Fig. 1.

The PatchMan protocol: flowchart. The input is a receptor PDB file and a peptide sequence. 1) Definition of surface motifs on the receptor: the protein surface is defined based on solvent accessibility, and then split into small structural surface patches. 2) Identification of structural matches in protein structures: matches are detected using MASTER search against a nonredundant dataset of protein structures. 3) Generation of the peptide–protein complex structure: the peptide fragment is determined (see enlargement) and superimposed onto the receptor. 4) Threading, refinement and scoring: the peptide sequence is threaded onto the identified complementing fragment and the structure is refined using the Rosetta FlexPepDock refinement protocol. The generated structures are clustered, and top-scoring cluster representatives are selected as final predictions. Enlargement in step 3: Extracting peptide fragments. Neighboring residues (magenta) around the matching motif (green) are defined as Cβ distance within 8 Å of the motif. Consecutive backbone stretches are then elongated in both directions to the desired peptide length. Arrows indicate stretches that can be elongated. Single residue (indicated with dashed circle) will not be elongated. See Methods for more details.