On page 49 (DOI: 10.1002/jcc.23771), Chengfei Yan and Xiaoqin Zou describe a novel computational approach, ACCLUSTER, to predict peptide binding sites on protein surfaces by clustering chemical interactions. The method is assessed by diverse proteinpeptide complexes, yielding very good performance. The method does not involve any training database, and can be easily extended to other systems, such as RNA-peptide interactions, for which experimental structural data are insufficient for informatics-based modeling. The cover shows an example (pdb: 1CJF) in which ACCLUSTER correctly identifies the peptide binding site as the top prediction. The predicted sites ranked as numbers one, two, and three are colored red, green, and magenta, respectively.