An efficient data filtering strategy for easy metabolite detection from the direct analysis of a biological fluid using Fourier transform mass spectrometry

High-throughput analyses require an overall analytical workflow including not only a robust and high-speed technical platform, but also dedicated data-processing tools able to extract the relevant information. This work aimed at evaluating post-acquisition data-mining tools for selective extraction of metabolite species from direct introduction high-resolution mass spectrometry data. Investigations were performed on spectral data in which seven metabolites of vinclozolin, a dicarboximide fungicide containing two chloride atoms, were previously manually identified. The spectral data obtained from direct introduction (DI) and high-resolution mass spectrometry (HRMS) detection were post-processed by plotting the mass defect profiles and applying various data-filtering methods based on accurate mass values. Exploration of mass defect profiles highlighted, in a specific plotting region, the presence of compounds containing common chemical elements and pairs of conjugated and non-conjugated metabolites resulting from classical metabolic pathways. Additionally, the judicious application of mass defect and/or isotope pattern filters removed many interfering ions from DI-HRMS data, greatly facilitating the detection of vinclozolin metabolites. Compared with previous results obtained by manual data treatment, three additional metabolites of vinclozolin were detected and putatively annotated. Tracking simultaneously several specific species could be efficiently performed using data-mining tools based on accurate mass values. The selectivity of the data extraction was improved when the isotope filter was used for halogenated compounds, facilitating metabolite ion detection even for low-abundance species