Related Experiment Video
Updated: Oct 29, 2025
10:10
Genetic Engineering of an Unconventional Yeast for Renewable Biofuel and Biochemical Production
Published on: September 20, 2016
14.5K
Integrated knowledge mining, genome-scale modeling, and machine learning for predicting Yarrowia lipolytica
Jeffrey J Czajka1, Tolutola Oyetunde2, Yinjie J Tang1
1Department of Energy, Environmental and Chemical Engineering, Washington University, St. Louis, MO, 63130, USA.
Metabolic Engineering
|July 9, 2021
Summary
This study developed a machine learning model to predict bioproduction titers in Yarrowia lipolytica, integrating genome-scale modeling and experimental data. The model accurately predicts high titers (>1 g/L) and identifies key metabolic factors for improving microbial chemical production.
Area of Science:
- Biotechnology
- Metabolic Engineering
- Computational Biology
Background:
- Predicting microbial bioproduction titers is complex due to intricate regulatory networks and cultivation conditions.
- Accurate titer prediction is crucial for optimizing industrial microbial chemical production.
Purpose of the Study:
- To develop a predictive model for Yarrowia lipolytica chemical titers using integrated data.
- To identify key metabolic and genetic features influencing bioproduction performance.
Main Methods:
- Literature data mining and feature extraction.
- Genome-scale modeling (GSM) and Flux Balance Analysis (FBA) for metabolic flux estimation.
- Machine learning (ML) ensemble learner for titer prediction.
Main Results:
- Accurate prediction of Y. lipolytica titers >1 g/L (R² = 0.87).
- Reduced predictability for low-titer strains (<1 g/L, R² = 0.29), suggesting potential biosynthesis bottlenecks.
- Identified FBA fluxes, enzyme steps, substrate inputs, and thermodynamic barriers as key predictive features.
Conclusions:
- The integrated GSM-ML approach provides a powerful platform for predicting microbial titers.
- Conserved features across oleaginous yeasts suggest potential for transfer learning.
- The model can guide computational strain design for enhanced bioproduction.
