For decades, the drug design community was concerned with deriving insights for their targets from a static picture using X-ray crystallography and pharmacophores. Likewise, computational methods for in silico studies of ligand binding rely heavily on docking, a static method. Still, computational methods such as molecular dynamics (MD) simulations were gaining prominence in the field of computational drug design. In consequence, as the computational power increased and the simulation technology matured, MD simulations became more and more critical for the drug design process. Thus, innately, there have to be deep conflicts between the current pool of models derived from a static mindset and insights derived from dynamic modelling. The recent introduction of structure-based ensemble- QSAR methods[1] and dynamic pharmacophores, termed Dynophores[2] served as a primer to bridge the gap between the two ways of modelling supramolecular complexes common in computational drug design. Recently, Dynophores replaced classical pharmacophores in a virtual screening study on the DNA topoisomerase IIa[3]. The subsequent work focuses on the development and the application of a new method- ology deriving models from MD simulations using lean project management techniques, Dynophores and Markov modelling. The newly developed methodology abandons the static picture entirely, and investigates three prominent drug targets. The first target is the ZIKA virus protease and serves as a benchmark against domain expert knowledge for designing de novo inhibitors[4]. The second target is the human cyclin dependent kinase 2 (CDK2) complexed with three different ligands[5,6]. The system serves as a benchmark to derive new insights from experimental data, as the mechanisms of action and inhibition of CDK2 are up to now not genuinely understood. The third system is the hepatitis C virus (HCV) NS3/4A protease shall serve as a test if the method can explain resistance mechanisms in prominent pathogens against data derived from standard experimental protocols. Four different supramolecular complexes of the NS3/4A protease containing the wild type and three different mutations, each complexed with vaniprevir[7,8] were investigated. For all three test systems, new insights were generated
