ACCLUSTER predicts peptide binding sites on the protein structure by clustering chemical interactions between amino acid probes and protein receptors.

Specifically, we use the 20 standard amino acids (ACCLUSTER-SI) or amino acids occurred in the sequence of the binding peptide (ACCLUSTER-SD) if the sequence of the bound peptide is provided as probes to globally scan the surface of a given protein. Modes of the amino acids that form good chemical interactions with the protein are kept and then clustered by the Density- Based Spatial Clustering of Applications with Noise (DBSCAN) method. These clusters are then ranked based on their sizes. The top clusters are considered as the predicted peptide binding sites.

The performance of ACCLUSTER on PROPEP (251 protein-peptide complexes):

The performance of ACCLUSTER on PeptiDB (103 protein-peptide complexes and 78 unbound or homologous protein structures):