Enhanced Nontarget Screening Algorithm Reveals Highly Abundant Chlorinated Azo Dye Compounds in House Dust

Steven Kutarna,Song Tang,Xiaojian Hu,Hui Peng
DOI: https://doi.org/10.1021/acs.est.0c06382
2021-03-15
Abstract:Humans spend 90% of their time indoors, but the majority of indoor pollutants remain unknown. In this study, a nontarget screening algorithm with reduced false discovery rates was developed to screen indoor pollutants using the Toxic Substances Control Act (TSCA) database. First, a putative lock mass algorithm was developed for post-acquisition calibration of Orbitrap mass spectra to sub-ppm mass accuracy. Then, a one-stop screening algorithm was developed by combining MS<sup>1</sup> spectra, isotopic peaks, retention time prediction, and <i>in silico</i> MS<sup>2</sup> spectra. A sufficient true positive rate (73%) and false discovery rate (5%) were achieved for the screening of halogenated compounds at a score cutoff of 0.28. Above this cutoff, 427 chemicals were detected from 24 house dust samples, including 39 chlorinated compounds. While some identified halogenated compounds (<i>e.g.</i>, triclosan) are well known, 18 previously unrecognized chlorinated azo dyes were detected with high abundance as the largest class of chlorinated compounds. Two chlorinated azo dyes were confirmed with authentic standards, but the two most abundant chlorinated azo dyes were missed by the algorithm due to the limited breadth of the TSCA database. These compounds were annotated as chlorinated analogues of Disperse Blue 373 and Disperse Violet 93 using the DIPIC-Frag method. This study revealed the presence of highly abundant chlorinated azo dyes in house dusts, highlighting their potential health risks in the indoor environment.The Supporting Information is available free of charge at <a class="ext-link" href="/doi/10.1021/acs.est.0c06382?goto=supporting-info">https://pubs.acs.org/doi/10.1021/acs.est.0c06382</a>.Sample treatment and instrumental conditions; details of PLMA calibration; details of automated nontarget screening algorithm; list of 47 putative lock masses; list of functions used for PLMA curve fitting; list of neutral loss groups used for scoring; 30 reference halogenated compounds; 39 tentatively identified chlorinated compounds; mass errors of compounds before and after mass calibration; workflow of PLMA; workflow of extracting isotope information from MS<sup>1</sup> and MS<sup>2</sup>; formula matching for DV93; match scores of detected compounds; ΔScore of halogenated reference standards; impacts of excluding individual score components; compound match scores grouped by functional use; peak abundance distribution of all compound matches; validation of DB79:2; MS<sup>2</sup> spectra of DV93 and Cl-DV93; and chromatograms of chlorinated and brominated azo dyes (<a class="ext-link" href="/doi/suppl/10.1021/acs.est.0c06382/suppl_file/es0c06382_si_001.pdf">PDF</a>)List of 427 compound matches from the house samples (<a class="ext-link" href="/doi/suppl/10.1021/acs.est.0c06382/suppl_file/es0c06382_si_002.xlsx">XLSX</a>)This article has not yet been cited by other publications.
environmental sciences,engineering, environmental
What problem does this paper attempt to address?