Multilaboratory Study of a Nontarget Data Acquisition for Target Analysis (nDATA) Workflow Using Ultrahigh-Performance Liquid Chromatography-High-Resolution Mass Spectrometry for the Screening of 1087 Pesticides in Fresh Fruits and Vegetables

A nontarget Data Acquisition for Target Analysis (nDATA)
workflow
was developed to screen pesticides in fresh produce based on ultrahigh-performance
liquid chromatography-high-resolution full scan mass spectrometry/variable
data-independent tandem mass spectrometry acquisition (LC-FS MS/vDIA
MSMS) and a pesticide database. The MSMS spectral library was generated
to create a database consisting of 1087 pesticides based on authentic
pesticide standards. The retention time (±0.5 min), precursor
ion (≤± 5 ppm), and product ions (≤± 5 ppm)
were extracted for each pesticide from LC-FS MS/data-dependent MSMS
acquisition (LC-FS MS/DDA MSMS). Mass accuracy criteria, along with
±0.1 min retention time tolerance, were used for the identification
of pesticides. Three laboratories evaluated and validated the nDATA
workflow to screen and identify pesticides from produce extracts (apples,
bananas, broccoli, carrots, grapes, lettuce, oranges, potatoes, strawberry,
and tomatoes) prepared by the Quick, Easy, Cheap, Effective, Rugged,
and Safe (QuEChERS) procedure. Of the 1087 pesticides evaluated, false-positive
rates were ≤5% for 977 pesticides in blank matrices and false-negative
rates were ≤5% for 921 and 985 pesticides in fortified matrices
at 10 and 100 μg/kg, respectively. False positives detected
were misidentified pesticides, incurred residues, or contaminants
possibly resulting from process or system contamination detected below
the threshold level of 10 μg/kg. False negatives were attributed
to pesticides that did not sufficiently ionize or fragment or had
poor stabilities and QuEChERS extraction efficiencies. Incurred residues
in archived produce samples (apple, Chinese broccoli, grape, kale,
kohlrabi, orange, pepper, strawberry, tomato, and turnip green) were
prepared using QuEChERS, evaluated by the nDATA workflow, and the
results were compared and confirmed, if possible, to targeted GC-MS/MS,
LC-MS/MS, and LC-FS MS/DDA MSMS methods. The three laboratories identified
25 parent pesticides at levels >10 μg/kg that were consistent
with findings from targeted procedures and discovered 10 different
metabolites that were not provided in the multiple reaction monitoring
method or inclusion list of the targeted procedures. GC-MS/MS identified
two pesticides, chlorothalonil and dacthal, and a possible chlorothalonil
metabolite, pentachlorobenzonitrile, that were not amenable to LC-low
or LC-high-resolution mass spectrometry analysis in produce samples.
To improve the identification quality, the nDATA workflow further
implemented quality control, operational, and processing measures
to reduce the number of false detects, and the data evaluation workload.
As demonstrated in this study, the validated nDATA workflow creates
new opportunities for chemical residues analysis, offering a potential
screening complement to targeted LC-MS/MS, GC-MS/MS, and nontargeted
methods for pesticides and other contaminants of interest.

靶向分析非靶向数据采集（nDATA）工作流，是基于超高效液相色谱-高分辨全扫描质谱/可变数据非依赖型串联质谱采集（LC-FS MS/vDIA MSMS）技术与农药数据库开发而来，用于筛查新鲜农产品中的农药残留。研究通过真品农药标准品构建了包含1087种农药的串联质谱光谱库，并以此建立对应农药数据库。针对每种农药，从超高效液相色谱-全扫描质谱/数据依赖型串联质谱采集（LC-FS MS/DDA MSMS）数据中提取其保留时间（±0.5 min）、前体离子（≤±5 ppm）与产物离子（≤±5 ppm）信息；以质量准确度准则辅以±0.1 min的保留时间容忍度，作为农药定性识别的依据。三家实验室对该nDATA工作流进行了评估与验证，该工作流可对经快速、简便、经济、高效、稳定、安全（QuEChERS）前处理方法制备的农产品提取物（包括苹果、香蕉、西兰花、胡萝卜、葡萄、生菜、橙子、土豆、草莓与番茄）中的农药进行筛查与定性识别。在受试的1087种农药中，空白基质中977种农药的假阳性率≤5%；在加标浓度为10 μg/kg与100 μg/kg的基质中，分别有921种与985种农药的假阴性率≤5%。检出的假阳性结果主要包括误定性农药、固有残留或污染物，此类信号多源于过程污染或系统污染，且浓度低于10 μg/kg的阈值水平。假阴性结果则多归因于农药电离或碎裂不充分、稳定性较差，或是QuEChERS前处理提取效率不佳。采用QuEChERS前处理方法制备存档农产品样品（包括苹果、芥兰、葡萄、羽衣甘蓝、苤蓝、橙子、辣椒、草莓、番茄与芜菁叶）中的固有残留样品，通过nDATA工作流进行分析，并尽可能将分析结果与靶向GC-MS/MS、LC-MS/MS及LC-FS MS/DDA MSMS方法的检测结果进行对比验证。三家实验室共检出25种母体农药，其浓度均高于10 μg/kg，且与靶向检测方法的结果一致；同时发现了10种未被靶向检测方法的多反应监测（multiple reaction monitoring，MRM）方法或目标物清单纳入的代谢物。靶向GC-MS/MS方法还检出了两种不适用于液相低分辨或高分辨质谱分析的农药：百菌清与敌草索，以及一种可能为百菌清代谢物的五氯苯甲腈。为提升定性识别质量，nDATA工作流进一步增设了质量控制、操作规范与数据处理措施，以减少假阳性检出数量并降低数据评估工作量。本研究结果表明，经过验证的nDATA工作流为化学残留分析提供了新的思路，可作为靶向LC-MS/MS、GC-MS/MS及非靶向农药与其他目标污染物检测方法的潜在筛查补充工具。