Enzymes have been engineered for everything from food to fuel, pharmaceuticals to consumer products. Our lab is using machine learning to advance enzyme engineering, finding improved protein sequences faster and with fewer resources by augmenting directed evolution with in silico screening. We are interested in a broad range of topics across supervised and unsupervised learning for protein engineering and how machine learning can be applied across all steps of the directed evolution cycle.

Directed evolution is a process for engineering proteins inspired by artificial selection. It starts with an initial sequence having a detectable amount of the activity or other properties of interest. Mutations are introduced to create a set of mutant genes, which are expressed in a host organism and screened for the properties of interest. The sequence with the highest such ‘fitness’ becomes the starting point for the next round and the cycle continues until a desired fitness is reached.