A Paired Database of Predicted and Experimental Protein Peptide Binding Information

Abstract

Abstract Peptides are important biomolecules, and their interactions with proteins make them useful in sensing and therapeutic applications. Computational peptide design methods can benefit from high-quality peptide-protein structures paired with thermodynamic data. The Predicted and Experimental Peptide Binding Information (PEPBI) database provides 329 predicted peptide-protein complexes, each based on an experimentally determined structure, with corresponding experimental measurements of changes in Gibbs free energy, enthalpy, and entropy. For each complex, 40 properties calculated using Rosetta’s Interface Analyzer are included. Complexes were selected for inclusion in PEPBI using eight stringent structural criteria, including peptide length (5–20 residues), structure resolution (≤2.0 Å), less than 30% sequence identity between complexes, and having a corresponding unbound protein structure in the Protein Data Bank with at least 90% sequence identity to the bound form with minimal changes in the binding pocket. PEPBI is expected to be of use for the development of computational methods for peptide design with desired binding properties to protein targets.