A method implemented in the form of a computer simulation code for evaluating the free energy of binding between polypeptide amino acid residues and one or more molecular fragment types is presented. The basis of the method is a novel weighted Metropolis Monte Carlo approach for sampling the grand canonical ensemble. By making use of the properties of the grand canonical ensemble, the affinity of fragments for binding in the vicinity of each protein residue can be efficiently computed. The binding volume associated to each fragment-residue pair is estimated on the basis of a simple proximity criteria, and a useful affinity mapping of the protein surface can be obtained in this way. The analysis of such data for various fragment types provides valuable information to help identify protein binding sites, as well as to identify key fragments used for building potential drug leads.