The overall objective of our work is to provide an understanding of enzyme function through studying their structures and dynamics. Among others, we analyze the physicochemical properties of active sites using three-dimensional point clouds or molecular docking calculations. To better understand protein stability and conformational flexibility, we run molecular dynamics simulations. Specific examples include monoglyceride lipases from different organisms, where we investigate the role of cap dynamics in substrate specificity in collaboration with Monika Oberer, or novel phenolic acid decarboxylases obtained by “ancestral sequence reconstruction”, where we aim at identifying hotspots responsible for protein stability. In addition, we design machine learning algorithms to predict protein properties from sequence and mutagenesis data.