Accelerating Confident Unknown Identification Using Agilent Revident LC/Q-TOF and MassHunter Explorer 2.0

Significance of the topic

Environmental water monitoring increasingly requires non-targeted workflows capable of revealing both regulated and emerging contaminants that targeted methods miss. High-resolution LC/Q-TOF mass spectrometry combined with chemometrics and spectral/structure databases bridges this gap by enabling comprehensive screening, confident elemental formula assignment, and structural proposal for unknowns while addressing the practical bottleneck of data overload and false positives.

Objectives and overview of the study

This application note demonstrates a streamlined, project-oriented workflow for unknown compound identification in a fortified water sample. Objectives were to reduce data complexity, prioritize true analyte features over background, perform automated library matching across multiple curated databases, and validate putative identifications using machine learning–assisted structure elucidation (SIRIUS with CSI:FingerID). The workflow emphasizes automation and traceability to decrease manual peak picking and improve confidence in assignments.

Used instrumentation

Key hardware and software used in the study are summarized below:

• UHPLC: Agilent 1290 Infinity III platform (high-speed pump, multisampler, multicolumn thermostat, InfinityLab Assist Hub).
• Column: ZORBAX Eclipse Plus C18, 2.1 × 100 mm, 1.8 µm; column temp 45 °C; injection volume 20 µL.
• Mobile phases: 10 mM ammonium formate pH 5 (A) and same in methanol (B); flow 0.45 mL/min with a rapid gradient to 99% B.
• Mass spectrometer: Agilent Revident Quadrupole Time-of-Flight LC/MS with Dual Jet Stream (AJS) ion source operating in positive ion mode.
• Acquisition: full-scan MS (m/z 50–500) and directed MS/MS; MS tolerance 5 ppm, MS/MS tolerance 15 ppm.
• Software: MassHunter Acquisition and Qualitative Analysis for data capture and BPC review; MassHunter Explorer 2.0 for feature extraction, chemometrics, and library matching; Agilent ChemVista to manage PCDL libraries; SIRIUS with CSI:FingerID for formula and structure confirmation.

Methodology

The workflow combined chromatographic comparison, automated feature extraction, chemometric filtering, library searching, and structure validation:

1. Direct injection of fortified water and control (20 µL) and acquisition by LC/Q-TOF in positive mode.
2. Base peak chromatogram (BPC) comparison to visualize distinctive peaks between fortified sample and control.
3. Automated feature extraction in MassHunter Explorer 2.0 produced 528 features (default environmental method with modifications: height filter 5,000 counts, RT tolerance ±0.02 min, isotope model common organic).
4. Chemometric filtering using fold-change analysis (FC > 10) narrowed candidates to 51 features (>90% reduction).
5. Retention-time matching between significant features and BPC peaks identified ions of interest for directed MS/MS acquisition.
6. Library matching against four PCDL databases (pesticides, EPA1699 water screening, extractables & leachables, and MoNA) with MS/MS matching thresholds (5 ppm MS, 15 ppm MS/MS).
7. Independent validation of putative matches using SIRIUS with CSI:FingerID to predict elemental formulas and rank candidate structures using fragmentation-based machine learning.

Main results and discussion

Key outcomes and interpretive points:

• Data reduction: feature count was reduced from 528 to 51 by applying FC > 10, focusing effort on features elevated in the fortified sample and removing many matrix/background features.
• Detection and identification: 20 pesticides were detected and putatively identified. Seventeen of 20 BPC-distinct peaks matched pesticide entries in PCDLs with MS/MS scores >81 and mass accuracy better than 1 ppm.
• Validation by SIRIUS: CSI:FingerID provided orthogonal confirmation of elemental composition and candidate structures, helping to remove false positives and prioritize correct identifications.
• False positives and limitations of BPC-only assessment: Four peaks that visually differed in BPC but lacked unique feature support were identified as false positives after chemometric and MS/MS confirmation, showing that BPC inspection alone can be misleading.
• Detection despite similar BPC: Three pesticides (diuron, linuron, metobromuron) were recovered from chemometrically significant features despite minimal BPC differences, illustrating that statistical prioritization can retrieve true analytes hidden by background similarity.
• Isomer differentiation: Two structural isomers (sebuthylazine and terbuthylazine) sharing identical m/z and MS/MS spectra were resolved by retention time, emphasizing the importance of chromatographic separation and retention information for isomer assignments.
• Analytical performance: Sensitivity spanned approximately 8.7 × 10^4 to 2.8 × 10^6 counts using the chosen injection volume; MS mass errors were consistently <1 ppm for reported pesticides.

Benefits and practical applications of the method

The integrated workflow delivers several practical advantages:

• Substantial time savings and reduced operator bias through automated feature extraction, guided project workflows, and chemometric filtering.
• Higher confidence identifications by combining accurate mass, MS/MS spectral matching, and ML-based structure validation.
• Improved prioritization of true environmental contaminants over matrix and background features, reducing false positives.
• Applicability across domains: environmental monitoring, food safety screening, extractables & leachables assessment, and biomarker discovery.

Future trends and potential uses

Prospective directions and opportunities informed by this work include:

• Deeper integration of machine learning across the workflow to further automate candidate ranking, reduce human review, and detect low-abundance or novel transformation products.
• Expansion of curated spectral and retention time libraries (including collision cross-section data) to strengthen isomer discrimination and reduce reliance on manual confirmation.
• Routine adoption of project-based, reproducible pipelines for regulatory and monitoring laboratories to support large-scale surveillance and longitudinal studies.
• Coupling non-targeted screening with quantitative follow-up methods (e.g., targeted LC-MS/MS) to move from detection to actionable concentration data for risk assessment.

Conclusion

This application note demonstrates a reproducible, scalable workflow that leverages high-resolution LC/Q-TOF, chemometric filtering, multi-database spectral matching, and ML-enabled structure confirmation to accelerate confident unknown identification in water samples. The strategy reduced feature complexity by over 90%, identified 20 pesticides with sub-ppm mass errors, and reduced false positives compared with BPC-only approaches. The combined platform and software approach is well suited for routine environmental screening and can be adapted for related analytical challenges in food safety and biomarker research.

References

1. Nieto S., Chen K., Alaimo C., Young T. Comprehensive Profiling of Environmental Contaminants in Surface Water Using High-Resolution GC/Q-TOF. Agilent Technologies Application Note, publication 5994-1371EN, 2019.
2. Sartain M., Bertram L., McEachran A., Pyke J., Hoffmann M. A. Drug Metabolite Identification with a Streamlined Software Workflow: Combining Agilent Revident Q-TOF LC/MS and MassHunter Explorer 2.0. Agilent Technologies Application Note, publication 5994-8867EN, 2026.
3. Dührkop K., Fleischauer M., Ludwig M., Aksenov A. A., Melnik A. V., Meusel M., Dorrestein P. C., Rousu J., Böcker S. SIRIUS 4: A Rapid Tool for Turning Tandem Mass Spectra into Metabolite Structure Information. Nature Methods 2019, 16(4), 299–302.
4. Dührkop K., Shen H., Meusel M., Rousu J., Böcker S. Searching Molecular Structure Databases with Tandem Mass Spectra Using CSI:FingerID. Proceedings of the National Academy of Sciences USA 2015, 112(41), 12580–12585.
5. Kornas P., Chadha M. Quantitation of 764 Pesticide Residues in Tomato by LC/MS According to SANTE 11312/2021 Guidelines. Agilent Technologies Application Note, publication 5994-5847EN, 2023.