Metabolomics: Intelligent data acquisition for untargeted metabolomics followed by high-throughput quantitative metabolomics utilizing high-resolution accurate mass measurements

Importance of the Topic

Untargeted metabolomics aims to capture the widest possible range of small molecules in complex biological samples. Coupling this discovery approach with high-throughput targeted quantitation enables rapid identification and measurement of key biomarkers. In dairy and plant-based milk products, this integrated workflow supports quality screening, authentication, and nutritional profiling, matching industry needs for robust, reproducible assays and deeper metabolome coverage.

Objectives and Study Overview

This study combined an intelligent data acquisition strategy for untargeted liquid chromatography-mass spectrometry (LC-MS) with a streamlined targeted workflow for rapid amino acid and organic acid quantitation. The goals were to discover discriminating milk metabolite markers, develop a high-throughput targeted assay, validate method performance, and demonstrate classification of bovine versus plant-based and organic versus non-organic milk types.

Methodology

Sample preparation used a modified Folch extraction with chloroform-methanol-water and isotope-labeled internal standards. Two workflows were implemented:

• Untargeted analysis: A 15-minute reversed-phase gradient on C18 UHPLC with full scan, polarity switching (ESI+ and ESI−), and AcquireX deep-scan intelligent data acquisition.
• Targeted screening: Two 5-minute LC-MS methods for amino acids (ESI+) and organic acids (ESI−) using a simplified gradient.

Data processing employed Compound Discoverer for feature extraction and annotation, and TraceFinder for quantitation with internal calibration curves.

Used Instrumentation

• UHPLC: Thermo Scientific Vanquish Horizon system with Hypersil GOLD C18 column.
• Mass spectrometer: Thermo Scientific Orbitrap Exploris 240 with heated ESI probe and polarity switching.
• Software: Xcalibur, Compound Discoverer 3.3, TraceFinder 5.1.

Main Results and Discussion

AcquireX increased fragmentation coverage while reducing background spectra, improving annotation confidence across milk types. Validation of the targeted method showed linear calibration (R2 > 0.99), limits of quantitation down to sub-micromolar levels, and precision (%CV ≤ 10%). Multivariate analysis (PCA) distinguished bovine milks by fat content and organic status, and clearly separated plant-based samples. Key discriminant metabolites included amino acids (phenylalanine, isoleucine, leucine, valine, proline, alanine) and organic acids (maleic, succinic, gluconic, hippuric, orotic acids). Hippuric and orotic acids marked bovine milk, whereas gluconic acid was characteristic of soy milk.

Benefits and Practical Applications

This integrated approach delivers high data quality, robust quantitation, and confident compound identification. It supports rapid screening of milk authenticity, nutritional profiling, and quality control in dairy and plant-based beverage industries. The workflow’s throughput and sensitivity make it suitable for routine QA/QC and research applications.

Future Trends and Opportunities

Further integration of real-time intelligent acquisition, expanded spectral libraries, and advanced machine learning for annotation will enhance untargeted coverage. Miniaturized, multiplexed UHPLC-MS platforms and automated data processing pipelines will drive next-generation metabolomics in food analysis and clinical diagnostics.

Conclusion

The combined untargeted AcquireX and high-throughput targeted LC-MS workflow provides a powerful strategy for metabolite discovery and quantitation in complex dairy and plant-based matrices. It achieves deep metabolome coverage, high confidence annotation, and reliable quantitation, enabling rapid quality screening and authentication of milk products.