Suspect screening workflow comparison for the analysis of organic xenobiotics in environmental water samples

Suspect screening techniques are able to determine a broader range of compounds than traditional target analysis. However, the performance of the suspect techniques relies on the procedures implemented for peak annotation and for this, the list of potential candidates is clearly a limiting factor. In order to study this effect on the number of compounds annotated in environmental water samples, a method was validated in terms of absolute recoveries, limits of quantification and identification, as well as the peak picking capability of the software (Compound Discoverer 2.1) using a target list of 178 xenobiotics. Four suspect screening workflows using different suspect lists were compared: (i) the Stoffident list, (ii) all the NORMAN lists, (iii) suspects containing C, H, O, N, S, P, F or Cl in their molecular formula with more than 10 references in Chemspider and (iv) the mzCloud library. The results were compared in terms of the number of annotated compounds at each confidence level. The same 8 compounds (atenolol, caffeine, caprolactam, carbendazim, cotinine, diclofenac, propyphenazone and trimetoprim) were annotated at the highest confidence level using the four workflows. Remarkable differences were observed for lower confidence levels but only 4 features were annotated at different levels by the four workflows. While the third approach provided the highest number of annotated features, the workflow based on the mzCloud library rendered satisfactory results with a simpler approach. Finally, this latter approach was extended to the analysis of organic xenobiotics in different environmental water samples. (C) 2021 Elsevier Ltd. All rights reserved.

Automated summary

Suspect screening techniques can determine a wider range of compounds than traditional target analysis, but the performance relies on peak annotation procedures. A study compared four workflows using different suspect lists, revealing that while differences were observed at lower confidence levels, only 4 features were annotated differently among the workflows. The approach based on the mzCloud library provided satisfactory results with a simpler method for analyzing organic xenobiotics in environmental water samples.