structure-based drug design in cases where side-chain conformations are uncertain, be it through flexibility or the devised modeling procedure. The first approach is the knowledge-based scoring function ROTA, which has been derived to identify correct rotamers and near-native ligand placements. The second approach is the side-chain prediction algorithm IRECS, which has been developed for generating protein models that contain ensembles of rotamers for flexible side chains. Test results of both approaches from a virtual screening scenario will be shown, as also a molecular explanation of HCV becoming resistant under drug treatment, which was found with the help of IRECS.