Enzymes are Nature’s catalysts, reducing the timescales of the chemical reactions that drive life from millions of years to seconds. There is also great scope for enzymes as biocatalysts outside the cell, from therapeutic and synthetic applications, to bioremediation and even for the generation of novel biofuels. Recent years have seen several impressive breakthroughs in the design of artificial enzymes, particularly through experimental studies that iteratively introduce random mutations to refine existing systems until a property of interest is observed (directed evolution), as well as examples of de novo enzyme design using combined in silico / in vitro approaches. However, the tremendous catalytic proficiencies of naturally occurring enzymes are, as yet, unmatched by any man made system, in no small part due the vastness of the sequence space that needs navigating and the almost surgical precision by which enzymatic catalysis is regulated. The proposed work aims to combine state of the art computational approaches capable of consistently reproducing the catalytic activities of both wild-type and mutant enzymes with novel screening approaches for predicting mutation hotspots, in order to redesign selected showcase systems. Specifically, we aim to (1) map catalytic promiscuity in the alkaline phosphatase superfamily, using the existing multifunctionality of these enzymes as a training set for the introduction of novel functionality, and (2) computationally design enantioselective enzymes, a problem which is of particular importance to the pharmaceutical industry due to the role of chirality in drug efficacy. The resulting theoretical constructs will be subjected to rigorous testing by our collaborators, providing a feedback loop for further design effort and methodology development. In this way, we plan to push existing theoretical tools to the limit in order to bridge the gap that exists between the catalytic proficiencies of biological and man-made catalysts.