Complementary evaluation of computational methods for predicting single residue effects on peptide binding specificities

Understanding the interactions that make up protein-protein or protein-peptide interfaces is a crucial step towards applications in biotechnology. The ability to discriminate between different partners defines the specificity of a binding protein and is equally important as its affinity to the target. Whereas many established computational methods provide an estimate of binding or non-binding, comparing similar ligands is still significantly more challenging. Here we evaluated the capability of predicting ligand binding specificity using three established but conceptually different physics-based methods for protein design. As a model system, we analyzed the binding of peptides to designed armadillo repeat proteins, where a single residue of the peptide was changed systematically, and compared the results with an experimental reference data set. The mutation of a single residue can have a strong impact on binding affinity and specificity, which is difficult to capture in sampling and scoring. We critically assessed the prediction accuracy of the computational methods and found that the prediction performance of each method is differently affected, suggesting the use of a complementary approach of the evaluated methods.