Hybrid experimental-machine learning framework for media optimization enhances antioxidant production in Micrococcus...

Area of Science:

• Microbiology
• Biotechnology
• Computational Biology

Background:

• Microbial antioxidant production is vital but difficult to optimize due to complex nutrient interactions.
• Systematic optimization of culture media for antioxidant yield is challenging.

Purpose of the Study:

• To develop and validate a hybrid experimental-computational framework for optimizing antioxidant production in Micrococcus endophyticus SS-1.
• To identify novel culture media formulations that enhance antioxidant potency.

Main Methods:

• Implemented a hybrid framework combining orthogonal array design, neural network surrogate modeling, and genetic algorithm optimization.
• Evaluated ten nutritional and cultivation factors using an L81 orthogonal array.
• Validated optimized media through 2,2-diphenyl-1-picrylhydrazyl radical-scavenging assays and dose-response experiments.

Main Results:

• The genetic algorithm-optimized medium significantly increased antioxidant activity compared to baseline and orthogonal array formulations.
• Experimentally validated optimized media demonstrated enhanced antioxidant potency, indicated by a reduced half-maximal effective concentration (EC50).

Conclusions:

• The hybrid experimental-computational approach is effective for optimizing microbial antioxidant production.
• The optimized medium enhances the intrinsic potency of antioxidants produced by Micrococcus endophyticus SS-1.