Intelligent data acquisition for untargeted metabolomics of milk samples coupled with quantitative high-resolution accurate mass data collection

Significance of the Topic

The comprehensive analysis of milk metabolites is crucial for ensuring food quality, safety, and authentication. Untargeted metabolomics workflows enable the discovery of biochemical markers that differentiate milk types, support quality control, and enhance consumer protection.

Objectives and Overview of the Study

This study aimed to develop an intelligent data acquisition workflow for untargeted metabolomics of milk, achieving deep metabolome coverage and confident compound annotation. The approach combined reversed-phase UHPLC with high-resolution mass spectrometry and iterative inclusion/exclusion strategies to profile bovine milk of varying fat content alongside almond, oat, coconut, and soy milks. Key goals included identifying marker compounds for downstream high-throughput screening and authentication applications.

Methodology and Instrumentation

• Sample Preparation: Modified Folch extraction using chloroform:methanol and water, followed by isotopic internal standards (13C6-adipic acid, 13C4-aspartic acid).
• UHPLC Conditions: Thermo Scientific Vanquish Horizon with Hypersil GOLD C18 column (2.1×150 mm, 1.9 µm) at 45 °C, gradient elution with 0.1% formic acid in water/methanol.
• Mass Spectrometry: Thermo Scientific Orbitrap Exploris 240 with heated ESI probe in polarity-switching mode (70–800 m/z, 120 000 resolution, sub-ppm mass accuracy).
• Data Acquisition: AcquireX Deep Scan workflow for dynamic exclusion/inclusion list generation, enhancing MS/MS coverage of relevant metabolites.
• Data Analysis: Thermo Scientific Xcalibur and Compound Discoverer 3.3 for peak detection, compound identification, and multivariate analysis.

Main Results and Discussion

• AcquireX Workflow: Increased fragmentation of target metabolites while reducing background ions, improving annotation confidence over a wide dynamic range.
• Analytical Performance: Sub-ppm mass accuracy and low retention time and peak area CVs (< 5%) demonstrated high method robustness.
• Metabolic Profiling: Plant-based milks exhibited higher levels of amino acids (phenylalanine, isoleucine, leucine, valine, proline), whereas bovine milks contained elevated organic acids (maleic, succinic, gluconic acids).
• Multivariate Analysis: PCA clearly separated milk samples by fat content, origin (bovine vs. plant-based), and organic status, identifying discrimination markers for each group.

Benefits and Practical Applications of the Method

The optimized workflow offers a robust platform for deep metabolome interrogation of milk. It enables high-throughput screening assays for milk authentication, quality control, and differentiation of production methods (e.g., organic vs. non-organic).

Future Trends and Potential Applications

• Integration with machine learning for automated metabolite annotation and pattern recognition.
• Expansion to other dairy and non-dairy matrices to build comprehensive food metabolome libraries.
• Correlation of metabolic fingerprints with nutritional value, shelf life, and safety parameters.
• Standardization of workflows for regulatory and industry adoption.

Conclusion

An end-to-end untargeted metabolomics workflow was established using intelligent data acquisition and high-resolution mass spectrometry, enabling confident annotation and discrimination of milk types. Identified marker compounds lay the foundation for targeted screening assays that support food security and consumer protection.

References

• No external literature references were provided in the source document.