With computational protein folding methods there are approximately 600 million protein structures available, most not having an assigning function. We work on assigning and confirming protein functions for these solved structures.

Enzymes can be engineered (modified) to help them be better at the chemistry the catalyze, or can be programmed to have new functions. We are interested in how to improve or reprogram enzymes in order to solve real-world problems like improving biofuel production. 

We are interested in how machine learning can be used to streamline biochemistry research. Examples include predicting protein function and identifying optimal enzyme activity conditions.