A Comprehensive Untargeted Metabolomics LC/Q-TOF Workflow with an Unknowns Identification Strategy to Identify Plasma Metabolite Shifts in a Mouse Model

Significance of the topic

Untargeted metabolomics enables comprehensive profiling of endogenous small molecules to reveal biological perturbations and potential biomarkers. Its application in disease models supports mechanistic insights, therapeutic evaluation, and translational research.

Study objectives and overview

This work describes a fully integrated untargeted LC/Q-TOF metabolomics workflow for identifying plasma metabolite shifts in an obesity mouse model (DIO C57BL/6J). The pipeline spans automated sample preparation, robust HILIC chromatography, high-resolution mass spectrometry data acquisition, statistical analysis, and unknown compound identification.

Methodology

• Sample preparation: 20 µL mouse plasma processed via automated Captiva EMR-Lipid SPE on the Agilent Bravo Metabolomics Platform for high recovery and reproducibility.
• Chromatography: HILIC-Z method on an Agilent 1290 Infinity II Bio LC, optimized for retention time stability (RSD < 5% over 11 days) and inter-laboratory transferability.
• Mass spectrometry: Agilent 6546 LC/Q-TOF with Agilent Jet Stream source, All Ions acquisition at 6 Hz across three collision energies, delivering high resolution, isotopic fidelity, and extended dynamic range.
• Data processing: Feature extraction with MassHunter Profinder; statistical filtering (t-test p<0.05, fold change > 5, abundance > 5000) in Mass Profiler Professional; identification via embedded IDBrowser using custom HILIC PCDL and METLIN PCDL.
• Unknown identification: Targeted MS/MS reinjection of remaining features and structural annotation using SIRIUS and CANOPUS for class prediction and database linkage.

Instrumentation used

• Agilent Bravo Metabolomics Platform
• Agilent 1290 Infinity II Bio LC (HILIC-Z column)
• Agilent 6546 LC/Q-TOF with Jet Stream source

Main results and discussion

• High reproducibility confirmed: RT delta 0.02 min, mass accuracy < 0.5 ppm for internal standards.
• 2146 features extracted; statistical analysis yielded 47 significantly dysregulated metabolites between obesity and control mice.
• PCA and hierarchical clustering demonstrated clear group separation and metabolic pattern differences.
• 30 of 47 features annotated via custom PCDL; SIRIUS resolved 5 additional unknowns with high confidence (score ≥ 82), supported by CANOPUS classification.

Benefits and practical applications

The described workflow offers automated, low-volume plasma processing, robust chromatography transferable across labs, and reliable identification of polar metabolites. It accelerates biomarker discovery, supports QA/QC in clinical studies, and enhances reproducibility in metabolomics research.

Future trends and opportunities

Advancements may include expanded spectral and retention time libraries for deeper annotation, integration of ion mobility or single-cell metabolomics, and machine learning algorithms for feature prioritization. Scaling the workflow to larger cohorts and diverse biofluids will broaden its applicability.

Conclusion

A comprehensive untargeted metabolomics workflow combining automated SPE, robust HILIC-Z chromatography, high-resolution LC/Q-TOF acquisition, and iterative identification tools (IDBrowser, SIRIUS) was validated in a mouse obesity model. This platform offers reproducible, high-confidence metabolite characterization for preclinical and clinical applications.

References

• Sartain M, et al. Enabling Automated, Low-Volume Plasma Metabolite Extraction with the Agilent Bravo Platform. Agilent Technologies, 2020.
• Yannell KE, et al. Improvements to HILIC Robustness – a Targeted HILIC Metabolomics Method for Routine Analysis. ASMS, 2021.
• Dührkop K, et al. SIRIUS4: A Rapid Tool for Turning Tandem Mass Spectra into Metabolite Structure Information. Nat Methods, 2019.
• Dührkop K, et al. Systematic Classification of Unknown Metabolites Using High-Resolution Fragmentation Mass Spectra. Nat Biotechnol, 2020.