Machine-learning-guided directed evolution for protein engineering.

Area of Science:

• Biochemistry
• Computational Biology
• Protein Engineering

Background:

• Machine learning (ML) advances protein engineering by predicting sequence-function relationships without complex biophysical models.
• ML accelerates directed evolution by learning from variant data to select sequences with improved properties.

Purpose of the Study:

• To outline the methodology for constructing ML sequence-function models for protein engineering.
• To provide guidance on utilizing these models to direct protein engineering efforts.
• To review current ML applications and future potential in protein engineering.

Main Methods:

• Developing ML models to predict protein sequence-function relationships.
• Utilizing ML predictions to guide the selection of protein variants for directed evolution.
• Illustrating the process with two detailed case studies.

Main Results:

• Demonstrated the practical steps for building and applying ML sequence-function models.
• Showcased the acceleration of protein optimization through ML-guided directed evolution.
• Highlighted the potential for ML to discover novel protein functions.

Conclusions:

• ML-guided directed evolution is a powerful paradigm for optimizing protein functions.
• This approach enhances the efficiency and scope of protein engineering.
• Future ML applications promise to deepen our understanding of protein sequence-function landscapes.