Machine learning-aided design of composite mycotoxin detoxifier material for animal feed

Area of Science:

• Materials Science
• Animal Nutrition
• Toxicology

Background:

• Developing feed additives requires balancing efficacy with safety, often relying on extensive experiments.
• Mycotoxin contamination in animal feed poses significant health risks and economic losses.
• Current methods for mycotoxin detoxification can be costly and time-consuming.

Purpose of the Study:

• To present a novel approach for designing clay-based composite materials for mycotoxin removal from animal feed.
• To utilize machine learning to optimize material composition and understand mechanisms of action.
• To validate the efficacy of designed detoxifiers in vivo.

Main Methods:

• Machine learning models were trained on in vitro adsorption-desorption data.
• Computational screening identified optimal clay-based composite formulations.
• In vivo toxicokinetic studies in broilers assessed mycotoxin exposure biomarkers.

Main Results:

• Machine learning identified high-capacity mycotoxin removal formulations.
• A significant reduction (32% decrease in AUC for DON-sulphate) in deoxynivalenol (DON) systemic exposure was observed in broilers.
• The study provided insights into the mode of action of the clay-based detoxifiers.

Conclusions:

• Machine learning accelerates the design of effective clay-based mycotoxin detoxifiers for animal feed.
• The developed materials show promise in mitigating DON toxicity in poultry.
• This approach offers a cost-effective and efficient strategy for feed additive development.