Updated
Updated · Nature.com · Aug 3
REAP Delivers 104-Fold Enzyme Gain With AI-Robotics Platform
Updated
Updated · Nature.com · Aug 3

REAP Delivers 104-Fold Enzyme Gain With AI-Robotics Platform

1 articles · Updated · Nature.com · Aug 3

Summary

  • 104-fold enhancement in Staphylococcus aureus Sortase A marked the top result for REAP, a closed-loop AI and robotics system for automated enzyme engineering.
  • RankReg—a ranking-regression learning method at REAP's core—uses experimental feedback to navigate sparse data and rugged fitness landscapes, helping the platform find useful mutations in catalytic and distal regions.
  • Five cycles of testing also produced a 57-fold improvement in desired activity for cytochrome P450 BM3, showing the system can progress from single mutations to productive combinations.
  • The Nature Communications study presents REAP as a scalable, generalizable framework for faster biocatalyst discovery and broader enzyme-enabled synthesis.

Insights

Can an AI-driven robotic lab uncover hidden enzyme mutations that nature missed, revolutionizing how we manufacture drugs and chemicals?
Will the fusion of AI algorithms and robotic wet-labs render traditional protein engineering obsolete, or are we trusting black-box biology too soon?