Machine Learning-Guided Optimization of p-Coumaric Acid Production in Yeast.

Area of Science:

• Industrial Biotechnology
• Synthetic Biology
• Metabolic Engineering

Background:

• Microbial cell factories are crucial for a biobased economy.
• Design-Build-Test-Learn (DBTL) cycles accelerate development but require effective phase integration.

Purpose of the Study:

• To propose and evaluate integrated links within DBTL cycles for pathway optimization.
• To enhance p-coumaric acid (pCA) production in Saccharomyces cerevisiae.

Main Methods:

• Utilized one-pot library generation, random screening, and targeted sequencing.
• Employed machine learning (ML) models for pathway optimization and design space expansion.
• Applied feature importance and Shapley additive explanation values.

Main Results:

• Achieved a 68% increase in pCA production within two DBTL cycles.
• Reached a final titer of 0.52 g/L and a yield of 0.03 g/g on glucose.
• Demonstrated the robustness and flexibility of ML models in guiding optimization.

Conclusions:

• Integrated DBTL cycles with ML effectively accelerate microbial strain development.
• ML-derived insights expand the design space for improved biological production.
• This approach offers a robust strategy for optimizing biobased chemical production.