Optimizing Ethanol Production in Saccharomyces cerevisiae at Ambient and Elevated Temperatures through Machine...

Area of Science:

• Biotechnology
• Metabolic Engineering
• Synthetic Biology

Background:

• Growing demand for sustainable fuels drives interest in bioethanol.
• High-temperature fermentation is cost-effective but challenging for yeast.
• Saccharomyces cerevisiae strains exhibit poor fermentation at elevated temperatures.

Purpose of the Study:

• To optimize bioethanol production in Saccharomyces cerevisiae using a machine learning approach.
• To fine-tune promoter activities of endogenous genes for enhanced ethanol fermentation.
• To develop a cost-effective and efficient bioethanol production process.

Main Methods:

• Created 216 combinatorial yeast strains by replacing native promoters with varying strength promoters.
• Utilized XGBoost machine learning model trained on promoter strengths and metabolite data.
• Applied ML-guided workflow to optimize ethanol production at high temperatures (40 °C).

Main Results:

• Promoter replacement improved ethanol production by 63% at 30 °C.
• ML-guided optimization at 40 °C led to a 7.4% increase in ethanol yield.
• Developed a comprehensive library of promoter strength modifications for yeast engineering.

Conclusions:

• Machine learning effectively guides yeast strain optimization for bioethanol production.
• This approach accelerates metabolic engineering and enhances cost-effectiveness.
• Optimized yeast strains show potential for efficient, large-scale bioethanol manufacturing.