Fragment-based methods for drug optimization have great potential; however, time, expense and sensitivity considerations associated with NMR and X-ray crystallographic based methods limit their applicability. As an alternative we have developed a computational approach, SILCS: Site Identification by Ligand Competitive Saturation, that uses explicit solvent all-atom molecular dynamics to identify binding sites on protein surfaces for functional groups. Information from the SILCS approach may then be combined with structural information on an inhibitor-protein complex to facilitate modification of the ligand to improve its binding affinity. An overview of SILCS and its application to inhibitor-ligand optimization will be presented.

Cyclobutane pyrimidine dimers have been shown to be the most prevalent mutagenic photolesions found upon UV photoirradiation of DNA. Of these, the thymine dimer is often used as a model system for biophysical studies due to their high formation and equilibrium yields. We will show that nearest-neighbor approaches to evaluating sequence-dependent dimer formation yields are appropriate, and that these cannot simply be evaluated in terms of ground state conformational considerations, as evaluated using MD simulations.