Multilaboratory Collaborative Study of a Nontarget Data Acquisition for Target Analysis (nDATA) Workflow Using Liquid Chromatography-High-Resolution Accurate Mass Spectrometry for Pesticide Screening in Fruits and Vegetables

Nontarget data acquisition for target analysis (nDATA) workflows using liquid chromatography-high-resolution accurate mass (LC-HRAM) spectrometry, spectral screening software, and a compound database have generated interest because of their potential for screening of pesticides in foods. However, these procedures and particularly the instrument processing software need to be thoroughly evaluated before implementation in routine analysis. In this work, 25 laboratories participated in a collaborative study to evaluate an nDATA workflow on high moisture produce (apple, banana, broccoli, carrot, grape, lettuce, orange, potato, strawberry, and tomato). Samples were extracted in each laboratory by quick, easy, cheap, effective, rugged, and safe (QuEChERS), and data were acquired by ultrahigh-performance liquid chromatography (UHPLC) coupled to a high-resolution quadrupole Orbitrap (QOrbitrap) or quadrupole time-of-flight (QTOF) mass spectrometer operating in full-scan mass spectrometry (MS) data-independent tandem mass spectrometry (LC-FS MS/DIA MS/MS) acquisition mode. The nDATA workflow was evaluated using a restricted compound database with 51 pesticides and vendor processing software. Pesticide identifications were determined by retention time (t_R, ±0.5 min relative to the reference retention times used in the compound database) and mass errors (δ_M) of the precursor (RTP, δ_M ≤ ±5 ppm) and product ions (RTPI, δ_M ≤ ±10 ppm). The elution profiles of all 51 pesticides were within ±0.5 min among 24 of the participating laboratories. Successful screening was determined by false positive and false negative rates of <5% in unfortified (pesticide-free) and fortified (10 and 100 μg/kg) produce matrices. Pesticide responses were dependent on the pesticide, matrix, and instrument. The false negative rates were 0.7 and 0.1% at 10 and 100 μg/kg, respectively, and the false positive rate was 1.1% from results of the participating LC-HRAM platforms. Further evaluation was achieved by providing produce samples spiked with pesticides at concentrations blinded to the laboratories. Twenty-two of the 25 laboratories were successful in identifying all fortified pesticides (0-7 pesticides ranging from 5 to 50 μg/kg) for each produce sample (99.7% detection rate). These studies provide convincing evidence that the nDATA comprehensive approach broadens the screening capabilities of pesticide analyses and provide a platform with the potential to be easily extended to a larger number of other chemical residues and contaminants in foods.