WHAT’S IN THE DUST? GC◊GC-MS BASED NON-TARGET SCREENING OF HOUSE DUST

Significance of the Topic

Indoor environments contain complex mixtures of chemicals that accumulate in house dust, reflecting human activities, building materials and consumer products.
The presence of thousands of known and unknown compounds in dust poses potential health risks through ingestion, inhalation and dermal exposure.
Comprehensive non-target screening using advanced two-dimensional gas chromatography coupled to mass spectrometry (GC×GC-MS) enables high-resolution profiling of these mixtures.

Objectives and Study Overview

This study aimed to develop and validate an end-to-end experimental and data processing workflow for non-targeted GC×GC-MS analysis of house dust.
Key goals included:

• Optimization of sample extraction, clean-up and fractionation procedures.
• Evaluation of software tools and custom scripts for inter-sample data alignment, feature filtering and prioritization.
• Identification and tentative classification of chemicals of potential concern using spectral libraries and manual review.
• Implementation of semi-quantification strategies at multiple levels using internal and external standards.

House dust samples from Umeå (Sweden), Munich (Germany) and Lviv (Ukraine), as well as reference materials (EU composite and NIST SRM 2585) and laboratory blanks, were analyzed.

Methodology and Instrumentation

Sample Preparation and Fractionation:

• Ultrasonic extraction with hexane:acetone (3:1) and acetone.
• SPE clean-up using mixed-mode sorbents (PSA) with non-polar (dichloromethane) and polar (methanol:DCM) fractions.

GC×GC-MS Analysis:

• Instrumentation: LECO Pegasus BT 4D GC×GC-TOFMS.
• Chromatographic conditions optimized for broad chemical coverage.

Data Processing:

• Library search and initial classification with LECO ChromaTOF (v.5.51) and NIST23 spectral library.
• Custom data alignment using Python scripts and the Julia_Aligns_2DGC.jl package, incorporating filters for column bleed, zero-intensity and low-intensity ions.
• Spectral similarity algorithms (DISCO, NDP) for feature matching and post-alignment merging.
• Prioritization and (semi-)quantification using internal standards, reference standards and peak area comparisons.

Main Results and Discussion

Feature Detection and Classification:

• Over 3 100 library-matched features and approximately 6 500 unknown features detected across samples.
• Compounds spanned wide chemical classes such as phthalates, musks, surfactants, terpenes, pesticides and flame retardants.

Sample Comparison and Multivariate Analysis:

• PCA distinguished dust from different cities and reference materials, highlighting unique chemical fingerprints.
• Laboratory blanks and column bleed features were effectively removed by the filtering workflow.

Software Evaluation:

• Most off-the-shelf packages struggled with LECO .cdf output; only custom Python and Julia scripts provided robust alignment and filtering.
• Julia_Aligns_2DGC.jl achieved detection of 146 out of 152 spiked standards after manual review.

Benefits and Practical Applications

• Provides a streamlined, automated pipeline for high-throughput non-target GC×GC-MS data processing.
• Enhances detection of low-abundance and unknown compounds that may be overlooked by conventional GC-MS.
• Supports risk assessment by enabling semi-quantitative estimates and prioritization of potentially hazardous chemicals.
• Facilitates comparative studies of indoor chemical exposures across regions and building types.

Future Trends and Opportunities

• Integration of machine learning for automated feature annotation and toxicity prediction.
• Expansion of spectral libraries to include more emerging contaminants and transformation products.
• Development of inter-laboratory standards and shared data formats to improve reproducibility.
• Application of this workflow to other complex environmental matrices such as sediments and textiles.

Conclusions

A novel end-to-end workflow combining optimized sample preparation, advanced GC×GC-MS analysis and custom data alignment tools was demonstrated for non-target screening of house dust.
The approach revealed a vastly greater chemical complexity compared to traditional methods and provided a framework for prioritizing chemicals of concern.
Continued development of open-source data processing tools and expansion of spectral reference data will further enhance the capability to assess indoor chemical exposures.

References

• Gori T, et al. Basic Res Cardiol. 2020.
• Schweizer C, et al. J Expo Sci Environ Epidemiol. 2007.
• WHO. WHO Housing and Health Guidelines. 2022.
• REACH Registration Statistics. European Chemicals Agency. 2023.
• Haglund P, et al. Sci Total Environ. 2024.