Active Sites by Computational Protein Design

We have developed an automated method to design active sites into protein scaffolds using computational protein design techniques. We search through the amino acid sequence and conformation spaces by optimising protein stability and ligand binding. We use an all‐atom force field, a high‐ resolution protein structure and a rotamer library to model a protein’s unfolded and folded states. We enlarge a rotamer library by using a minimization procedure that optimizes rotamers to maximize intermolecular h‐bonds. We validate our methodology by re‐designing SH3‐domain proteins to bind a set of 64 peptides.