LC-MS/MS-based metabolomics coupled with machine learning for screening candidate biomarkers in bacon

Area of Science:

• Food Science
• Analytical Chemistry
• Computational Biology

Background:

• Traditional bacon authentication methods are subjective.
• Objective authentication is needed for quality control and consumer trust.

Purpose of the Study:

• To develop an objective discrimination framework for bacon authentication.
• To integrate metabolomics and machine learning for identifying discriminatory biomarkers.

Main Methods:

• Liquid chromatography-tandem mass spectrometry (LC-MS/MS) for metabolomic profiling.
• Machine learning algorithms (Random Forest, KNN, SVM, FNN) for feature selection and model validation.
• Statistical analysis (Kruskal-Wallis H test) to identify significant metabolites.

Main Results:

• Identified 100 differential metabolites between pork belly (LRW) and pork rump (LRT) bacon.
• Reduced to 22 key metabolites using feature selection techniques.
• Highlighted four candidate discriminatory metabolites: 6-hydroxyoctanoylcarnitine (LRT), dimethylethanolamine, xanthine, and ethyl hydrogen fumarate (LRW).
• Achieved robust discriminatory performance through bidirectional validation with machine learning models.

Conclusions:

• The developed framework offers an objective approach to bacon authentication.
• Identified specific metabolites can serve as biomarkers for bacon origin.
• This data-driven approach is generalizable for biomarker discovery in other meat products.