Accelerating large scale pesticide screening programs with HR LC–MS/MS using automated non-targeted analysis software

Significance of the topic

The detection and monitoring of pesticide residues in food and the environment remain central to public health, regulatory compliance and trade. Traditional targeted LC-MS/MS approaches (usually triple quadrupole) provide robust quantitation but are limited to predefined analyte lists, delaying response to emerging substances, metabolites and regulatory changes. High-resolution mass spectrometry (HRMS) combined with automated non-targeted analysis (NTA) software offers a streamlined route to broaden screening scope, integrate retrospective analysis and accelerate large-scale surveillance programs.

Objectives and study overview

This study evaluated a single HRMS LC–MS/MS acquisition workflow to support both targeted and non-targeted pesticide screening across multiple food commodities. The work focused on demonstrating:

• the feasibility of data-independent acquisition (DIA) HRMS for large-scale pesticide screening,
• the performance of an automated NTA software (Insight Profiler) for feature detection, alignment and identification using extensive suspect/spectral libraries, and
• the sensitivity and reliability of NTA versus conventional targeted triple quadrupole results, including detection of metabolites and compounds outside routine multiresidue panels.

Materials, methods and processing workflow

Samples: A set of unknown food commodity samples previously analyzed by triple quadrupole LC-MS/MS and QC samples spiked with a panel of pesticides were processed. A validation spike used olive oil matrix containing 280 pesticides at the maximum reporting limit (MRL), typically 10 µg/kg.

Chromatography and acquisition: Reversed-phase LC on a Shim-pack Velox Biphenyl (2.1 × 100 mm, 2.7 µm) at 40 °C with a 17-minute gradient; mobile phases contained 2 mM ammonium formate and 0.002% formic acid in water (A) and methanol (B). Injection included a 4 µL sample with 40 µL water co-injection.

Mass spectrometry: High-resolution QTOF (Shimadzu LCMS-9050) in positive electrospray ionization. TOF mass scan m/z 140–925 (100 ms) plus 31 DIA MS/MS windows (total cycle ~1 s). Precursor isolation widths: 20 Da for m/z 140–540 and 35 Da for m/z 540–925. Collision energy spread 5–55 V. External mass calibration applied.

Data processing: Insight Profiler automated workflow performed feature detection (low threshold), alignment across samples, application of statistical filters to remove high-variance ions and highlight significant signals, and library/suspect-list searching. Processing captures the full sequence of steps for batch analysis, enabling reproducible automated NTA.

Used instrumentation

• LC system: Shimadzu Nexera X2.
• Column: Shim-pack Velox Biphenyl (2.1 × 100 mm, 2.7 µm).
• HRMS: Shimadzu LCMS-9050 QTOF with DIA acquisition (31 variable windows).
• Software: Insight Profiler for feature detection, alignment and library-driven identification; third-party spectral libraries (MassBank, Food Safety Mass Spectral Library (Wageningen) and in-house pesticide library of ~500 entries).

Main results and discussion

Validation: Feature detection and library matching identified all 280 pesticides spiked into olive oil at 10 µg/kg, demonstrating sufficient sensitivity of the combined LC–QTOF acquisition and Insight Profiler workflow for a broad panel at regulatory-relevant concentrations.

Concordance with targeted methods: NTA processing of real food commodities confirmed compounds previously quantified by triple quadrupole LC-MS/MS (examples include acetamiprid, azoxystrobin, tebuconazole, etc.), showing agreement between targeted and HRMS NTA outputs.

Additional findings beyond targeted scope: NTA found residues not included in the routine multiresidue triple quadrupole panel. Notable examples:

• High pyrimethanil levels in nectarine—this compound was outside the targeted method scope, illustrating NTA’s ability to detect unexpected residues.
• Detection of dimethomorph in grape using the Food Safety Mass Spectral Library; this fungicide is banned in the EU as of 2024 but remains permitted in the UK until 2030, highlighting regulatory divergence and the need for broad screening.
• Identification of pesticide metabolites such as acetamiprid-N-desmethyl and propamocarb N-oxide, demonstrating the workflow’s capability to monitor biotransformation products relevant for food safety.

Utility of spectral libraries: Combining an in-house pesticide library (>500 entries) with third-party open libraries increased identification coverage and allowed recognition of residues not routinely targeted. The study emphasizes the value of extensive, high-quality HR-MS/MS libraries for reliable NTA identifications.

Limitations and considerations: While NTA provided comprehensive screening, final confirmation and accurate quantitation of novel or unexpected hits typically require standards and targeted follow-up. Matrix effects, library quality, and DIA co-isolation complexity can affect confidence levels and require careful interpretation and orthogonal confirmation when regulatory action is considered.

Benefits and practical applications of the method

• Single HRMS acquisition supports retrospective and prospective screening for both known and unknown residues, reducing the need for multiple targeted methods.
• Automated, configurable workflows (Insight Profiler) enable batch processing and reproducibility across large sample sets, improving throughput for surveillance labs.
• Capability to detect metabolites and compounds outside routine panels enhances food safety monitoring and helps laboratories adapt to changing regulatory landscapes.
• Integration with external spectral repositories expands identification potential without requiring exhaustive in-house library development.

Future trends and opportunities

• Expansion and curation of high-resolution spectral libraries (community-driven, open-access collections) will increase NTA confidence and reduce false positives.
• Improved DIA window strategies, ion-mobility integration and hybrid acquisition modes can reduce spectral complexity and improve MS/MS specificity for crowded samples.
• Machine learning and automated scoring systems can help triage and rank suspect identifications, accelerating expert review.
• Regulatory frameworks may evolve to accept HRMS-based screening as a first-line surveillance tool, with defined workflows for confirmation and quantification using targeted assays.
• Broader adoption of NTA will facilitate routine monitoring of metabolites, transformation products, and emerging contaminants across food and environmental matrices.

Conclusion

The study demonstrates that a single HRMS DIA LC–MS/MS method combined with automated non-targeted analysis software (Insight Profiler) can effectively screen large pesticide panels across diverse food matrices. The workflow identified all 280 pesticides in a validation spike at 10 µg/kg, matched targeted triple quadrupole results in real samples, and detected additional residues and metabolites outside routine multiresidue methods. Adoption of HRMS-based NTA workflows enhances surveillance breadth and flexibility, though confirmatory targeted analysis remains necessary for regulatory action.

References

1. Padilla-González F, Rizzo S, Dirks C, Bergkamp W, Rasker S, Aloisi I. Creation of an Open-Access High-Resolution Tandem Mass Spectral Library of 1000 Food Toxicants. Analytical Chemistry. 2025;97(43):23822–23830. DOI: 10.1021/acs.analchem.5c03020.