A Structural Comparison Approach for Identifying Small Variations in Binding Sites of Homologous Proteins

A method for analyzing the protein site similarity was devised aiming at understanding selectivity of homologous proteins and guiding the design of new drugs. The method is based on calculating Cα distances between selected pocket residues and subsequent analysis by multivariate methods. Five closely related serine proteases, the coagulation factors II, VII, IX, X, and XI, were studied and their pocket similarity was illustrated by PCA clustering. OPLS-DA was then applied to identify the residues responsible for the variation. By combining these two multivariate methods, we could successfully cluster the different proteases according to class and identify the important residues responsible for the observed variation.