Opportunities and Challenges for Machine Learning-Assisted Enzyme Engineering.

Area of Science:

• Biochemistry
• Computational Biology
• Protein Engineering

Background:

• Enzyme engineering optimizes protein properties like stability and efficiency through amino acid sequence modification.
• Traditional methods involve directed evolution, which can be time-consuming due to the vast protein search space.
• Machine learning (ML) is emerging as a complementary tool to empirical enzyme engineering.

Purpose of the Study:

• To explain how ML complements traditional enzyme engineering.
• To discuss the future potential of ML in advancing enzyme engineering outcomes.
• To highlight ML's role in both starting point discovery and fitness optimization.

Main Methods:

• ML models for functional annotation of known protein sequences.
• ML for generating novel protein sequences with desired functions.
• ML-based navigation of protein fitness landscapes by learning sequence-fitness relationships.

Main Results:

• ML aids in identifying suitable enzyme starting points.
• ML facilitates the optimization of enzyme fitness for specific applications.
• ML models can predict or generate sequences with enhanced properties.

Conclusions:

• ML significantly complements and accelerates enzyme engineering.
• ML offers powerful tools for discovering and optimizing enzymes with novel functions.
• The integration of ML promises to unlock improved engineering outcomes and expand enzyme capabilities.