Identifying the Sulfate Ion Binding Residues in Proteins

Many proteins function execution depends on the process of protein and ligand interact with each other. The identification of ligand binding residues is important for the research of the protein function. The 4442 protein chains with <25% sequence identity and resolution <3.0 angstrom were analyzed using Ligand Protein Contact database. Our final dataset contained 8112 sulfate ion binding residues (SIBR). We did a statistical analysis on window size as 7 amino acids. Using the amino acid composition, hydropathy information, correlation information and predicted structure information as the characteristic parameter, a Support Vector Machine algorithm for identifying sulphate ion binding residues was proposed. The overall accuracy and Matthew's correlation coefficient achieved 78.5% and 0.571 using the 5 fold cross validation. The Acc and MCC achieved 72.7% and 0.455 by using independent test. In addition, an online web server was established. http://202.207.30.72:7321/