A tool for calculating binding-site residues on proteins from PDB structures

In this study, we have developed a tool for calculating binding-site residues on proteins, TCBRP http://yanbioinformatics.cs.usu.edu:8080/ppbindingsubmit webcite. For an input protein, TCBRP can quickly find all binding-site residues on the protein by automatically combining the information obtained from all PDB structures that consist of the protein of interest. Additionally, TCBRP presents the binding-site residues in different categories according to the interaction type. TCBRP also allows researchers to set the definition of binding-site residues.The developed tool is very useful for the research on protein binding site analysis and prediction.Proteins perform various functions through interactions with other molecules, such as DNA, RNA, proteins, carbohydrates, and ligands. To understand the mechanisms of these interactions, many studies have been performed to analyze the properties of binding sites on proteins, such as residue composition, secondary structure, geometric shape, electrostatic potentials, etc [1-10]. To perform such an analysis, researchers must first identify the amino acid residues (referred to as binding-site residues) that are involved in the interactions. In other studies where the goal is to build computational predictors for predicting functional sites on proteins (e.g. DNA-binding sites, RNA-binding sites, protein-protein binding sites), researchers also need to identify binding-site residues on the training and test sets to train and evaluate their predictors [11-17].In most, if not all, of the above-mentioned studies, the binding-site residues are calculated from the 3-dimensional (3D) structures deposited in Protein Data Bank (PDB) [18]. Usually, researchers collected a non-redundant set of PDB structures, and then calculated binding-sites based on the PDB structures. However, one serious problem with this approach is that a protein may have multiple binding sites that interact with different interacting partners, but one PDB structure