An intelligent data acquisition workflow for untargeted metabolomics to achieve deep metabolome coverage and confident compound annotation

Importance of the topic

Untargeted metabolomics enables comprehensive profiling of small molecules in complex biological matrices. A robust workflow with deep metabolome coverage and confident compound annotation is critical for food quality assessment, authenticity verification, and consumer safety.

Aims and study overview

This study developed an intelligent data acquisition strategy using Deep Scan AcquireX on an Orbitrap Exploris 240 mass spectrometer to maximize MS/MS coverage in milk samples. The workflow was applied to bovine milk with varying fat content and plant-based alternatives (almond, oat, coconut, soy) to identify key metabolite differences for quality screening and authentication.

Methodology

A modified Folch extraction spiked with isotope-labeled adipic and aspartic acid standards ensured data quality. A 19-minute reversed-phase UHPLC separation on a Hypersil Gold C18 column preceded full scan (70–800 m/z) with polarity switching at 120 000 resolution. The AcquireX Deep Scan mode automatically generated exclusion and inclusion lists iteratively to enrich relevant MS/MS acquisitions.

Used Instrumentation

• Thermo Scientific Vanquish Horizon UHPLC system with autosampler at 5 °C
• Hypersil Gold C18 column (2.1 × 150 mm, 1.9 µm) at 45 °C
• Thermo Scientific Orbitrap Exploris 240 mass spectrometer with heated ESI probe and polarity switching
• TurboVap LV nitrogen evaporator
• Thermo Scientific Xcalibur and Compound Discoverer 3.3 software

Main results and discussion

The Deep Scan AcquireX workflow increased the proportion of fragmented metabolites while reducing background interference, enabling confident annotation across a wider dynamic range. Method validation showed sub-ppm mass accuracy, retention time CV below 0.5%, and stable signal response. PCA revealed clear separation by fat content in bovine milk and distinct clustering between organic and non-organic samples. Plant-based milks exhibited higher amino acid levels, whereas bovine milks contained elevated organic acids.

Benefits and practical applications

The optimized workflow offers high sensitivity, reproducibility, and extensive coverage for untargeted metabolomics in dairy matrices. Identified marker compounds support targeted high-throughput screening for milk authentication, quality control, and enhanced food security.

Future trends and potential applications

Upcoming advances may include machine learning–driven annotation, expansion to other food and environmental samples, and integration with targeted quantitation methods. Enhanced acquisition strategies will further improve depth of coverage and throughput.

Conclusion

An end-to-end untargeted metabolomics pipeline leveraging intelligent acquisition on high-resolution Orbitrap instrumentation enables deep metabolome coverage and reliable compound identification in milk. This approach lays the groundwork for targeted screening assays to ensure product authenticity and consumer protection.