The goal of this project is to develop and apply computational methods to reproduce, understand and predict structural and thermodynamical aspects of molecular recognition in the EcoRV enzyme-DNA and the small ligand-DNA complexes. We will use this methodology with a protein (EcoRV) interacting with DNA and small ligands: actinomycin, acridine, netropsin, furimidazoline. We will use the available experimental data on these systems as a critical guideline to evaluate the theory. Understanding recognition in the systems proposed here, especially in the EcoRV-DNA complex, will further the long term goal of the parent NIH grant, which is to make the computer approaches reliable and truly predictive in order to use them as an important tool in understanding molecular interactions as well as in the drug design process.