An Automated Toxicity Based Prioritization Framework for Fast Chemical Characterization in Non-Targeted Analysis

Junjie Yang,Fanrong Zhao,Jie Zheng,Yulan Wang,Xunchang Fei,Yongjun Xiao,Mingliang Fang
DOI: https://doi.org/10.1016/j.jhazmat.2023.130893
IF: 13.6
2023-01-01
Journal of Hazardous Materials
Abstract:Identification of environmental pollutants with harmful effects is commonly conducted by non-targeted analysis (NTA) using liquid chromatography coupled with high-resolution mass spectrometry. Prioritization of possible candidates is important yet challenging because of the large number of candidates from MS acquisitions. We aimed to prioritize candidates to the exposure potential of organic chemicals by their toxicity and identification evidence in the matrix. We have developed an R package application, “NTAprioritization.R”, for fast prioritization of suspect lists. In this workflow, the identification levels of candidates were first rated according to spectral matching and retention time prediction. The toxicity levels were rated according to candidates’ toxicity of different endpoints or ToxPi score. Finally, the various levels of candidates were identified as Tier 1 – 5 descending in priority. For validation, we used this workflow to identify pollutants in a sludge water sample spiked with 28 environmental pollutants. The workflow reduced the candidate list of over 6,982 candidates to a final list of 2,779 compounds and prioritized them to 5 tiers (Tier 1 – 5), including 21 out of 28 spiked standards. Overall, this study shows the added value of an automated prioritization R package for the fast screening of environmental pollutants based on the NTA method.
What problem does this paper attempt to address?