Peptides have recently attracted much attention as promising drug candidates. Rational design of peptide-derived therapeutics usually requires structural characterization of the underlying protein peptide interaction. Given that experimental characterization can be difficult, reliable computational tools are needed. In recent years, a variety of approaches have been developed for `protein-peptide docking', that is, predicting the structure of the protein-peptide complex, starting from the protein structure and the peptide sequence, including variable degrees of information about the peptide binding site and/or conformation. In this review, we provide an overview of protein-peptide docking methods and outline their capabilities, limitations, and applications in structure-based drug design. Key challenges are also briefly discussed, such as modeling of large-scale conformational changes upon binding, scoring of predicted models, and optimal inclusion of varied types of experimental data and theoretical predictions into an integrative modeling process.
