Related Experiment Video
Updated: Jan 18, 2026
Workflow Based on the Combination of Isotopic Tracer Experiments to Investigate Microbial Metabolism of Multiple Nutrient Sources
Published on: January 22, 2018
Artificial intelligence-driven fermentation optimization for α-amylase hyperproduction enabled by Raman monitoring
Yuan Wang1, Yonghong Wang1, Feng Xu2
1State Key Laboratory of Bioreactor Engineering, Qingdao Innovation Institute of East China University of Science and Technology, East China University of Science and Technology, Shanghai 200237, China; National Center of Bio-Engineering & Technology (Shanghai), East China University of Science and Technology, Shanghai 200237, China.
Abstract:
α-Amylase is a high-value enzyme widely applied in food, feed, textile, and bioenergy industries, yet achieving stable high-level production in Aspergillus niger remains difficult due to nonlinear fermentation dynamics and limited real-time control. To this end, an AI-driven fermentation optimization framework was established by combining multivariate machine learning, Raman spectroscopy-based glucose monitoring, and time-series transcriptomics. Twelve algorithms were benchmarked, with Random Forest showing the best predictive power, while SHAP interpretation highlighted glucose as the dominant regulatory factor. Prospective validation across different yield levels confirmed the model's robustness. Moreover, closed-loop Raman feedback maintained glucose within the optimal range, increasing α-amylase yield by 46 % and reducing fermentation time by 28 h compared with manual feeding. The optimized process achieved the highest α-amylase yield reported to date (15,729.47 U/mL). Overall, the integrative framework provides a scalable, mechanistic, and data-driven strategy for intelligent fungal fermentation and can be extended to other microbial bioprocesses.
