Current Research

Fully Autonomous Engineering of Proteins 

I work to combine machine learning techniques with a robotic experimental platform. The machine learning component designs experiments based on a library of possible protein sequences and instructs the robot to perform those experiments. When complete, the robot reports the result of the experiment back to the machine learning component, which designs follow-up experiments based on the new information. At no point in the process is a researcher required to influence experiment design or perform any experimental tasks, so the system can run continuously until the engineering goal is complete. For our proof-of-concept, we are working to optimize thermostability in a library of chimeric beta-glucosidase sequences.