Nuria Plattner - Freie Universität Berlin Markov modeling in combination with high-throughput molecular dynamics simulations has been used to build models of protein-protein interactions in full atomic detail for the ribonuclease Barnase and its natural inhibitor Barstar. Furthermore the method has been used to analyze the conformational dynamics of both proteins individually. The proteins initially associate into a structurally diverse set of non-native encounter states. This ensemble eventually funnels into an intermediate transition state in which Barnase and Barstar are flexibly associated before reaching the final bound configuration which remains stable for minutes. The Formation of the protein complex affects the conformational dynamics of the two proteins in different ways: for Barstar, the conformational dynamics in the complex can be described by a subset of the conformational states of the unbound protein. For Barnase in contrast new states emerge in the complex which are not observed for the unbound protein. Implications of this change in conformational dynamics for coarse grained descriptions of protein interactions are outlined.