Finding one mutation that improves a protein is hard. Finding five that work together is exponentially harder.
Today in @science.org, Hsu and Konermann labs present MULTI-evolve, a lab-in-the-loop framework that does it in just one machine learning-guided round.
