Accelerating Directed Evolution with Machine Learning: Insights and Learnings

  • Leveraged AI/ML tools to engineer a KRED capable of dynamic kinetic resolution, achieving ~1,500-fold improvement in enzymatic performance relative to wild-type
  • Accelerated enzyme evolution using an ML-guided algorithm, requiring over an order of magnitude fewer variants screened compared to traditional directed evolution
  • Conducted a retrospective analysis of ML programs applicable to biocatalysis using data from this campaign; key learnings will be shared