Developing new enzymes for catalyzing non natural reactions is a grand challenge for obtaining more efficient synthetic strategies for chemical industry as well as building complex molecules under enantioselective control with potential applications in medicine. Given the high specificity of natural enzymes for their substrates, new techniques are necessary for redesigning natural enzymes to perform novel reactions. Recently the group led by Prof. David Baker at University of Washington in Seattle (outgoing host) has made a breakthrough in the development of Computational Enzyme Design strategies to create new enzymes able to catalyze retro-aldol, Kemp elimination and Diels-Alder reactions. The methodology they have developed has proved to be a powerful tool for obtaining new activities with rate-enhancements up to 104 that allow further optimization with directed evolution techniques. The catalytic efficiency of designed enzymes, however, is much lower than that of natural enzymes and significant improvements are required to extend the applicability of these methods to any desired reaction. The first goal of the current project is to improve the current computational enzyme design methodology with a description of protein dynamics in order to achieve catalytic efficiencies closer to that of natural enzymes. As a second goal, we aim to use the improved Enzyme Design methodology to improve the activity of past designs and create new enzyme catalysts for the Bayliss-Hillman reaction. This is a carbon-carbon bond formation reaction that creates a new chiral center that allows to increase the complexity of molecules with potential applications in biomedicine. To complement the design process, QM/MM calculations will be performed to evaluate the reactivity of different designs and give insight into possible ways of improving the activity.