Self-driving autonomous machines to engineer novel enzymes

Area of Science:

• Biotechnology
• Computational Biology
• Protein Engineering

Background:

• Designing proteins with specific attributes, like enhanced thermostability, is complex.
• Automated platforms can accelerate the discovery and engineering of novel proteins.

Purpose of the Study:

• To describe a prototype of a self-driven laboratory for autonomous protein engineering.
• To demonstrate the capability of an automated system to design an enzyme with a desired attribute.

Main Methods:

• Developed the Self-driving Autonomous Machines for Protein Landscape Exploration (SAMPLE) system, integrating computational agents and automated robotic platforms.
• Utilized a computational component to learn protein sequence-function relationships from data.
• Employed an automated robotic system for protein synthesis, experimental characterization, and feedback generation.

Main Results:

• The SAMPLE system successfully navigated the protein landscape to engineer a glycoside hydrolase with enhanced thermostability.
• Four autonomous agents independently converged on a thermostable enzyme, showcasing the system's robustness.
• The automated process required no human intervention to achieve the desired outcome.

Conclusions:

• Self-driven laboratories represent a paradigm shift in autonomous scientific discovery.
• The SAMPLE system demonstrates the potential for AI-driven platforms to accelerate enzyme engineering and other complex biological design tasks.