Key Points:
- Urban runoff is a complex and undercharacterized source of chemical pollution. It contains numerous emerging contaminants such as 6PPD-quinone, HMMM, and DPG, many of which originate from vehicle traffic and tire wear.
- A combined DDA-DIA mass spectrometry workflow significantly improves identification of unknown pollutants.
- The method enabled tentative identification of previously unreported tire-related compounds in runoff. New pollutants such as CPG and BBG were detected frequently and are likely derived from tire wear, highlighting the method's potential for discovering environmentally relevant micropollutants.
High-resolution mass spectrometry (HRMS) is commonly used for non-target screening (NTS) of environmental samples, but identifying unknown pollutants remains difficult because of complex sample matrices and limited reference data. This study presents a workflow combining iterative data-dependent acquisition (DDA) with data-independent acquisition (DIA) to enhance compound annotation. Pooled samples were analyzed using iterative DDA, annotated with Sirius, and matched to DIA features in individual samples. Of the 44 target compounds in runoff, 50% were correctly identified as the top Sirius hit, with an additional 18% recovered through manual review—outperforming earlier studies using single-injection DDA. The method tentatively identified several tire-related compounds in urban runoff, including two novel findings in environmental samples.
Urban runoff is a major source of chemical pollution, originating from various sources including tire wear and vehicle traffic (1). Emerging contaminants such as 2-anilino-5-(4-methylpentan-2-ylamino)cyclohexa-2,5-diene-1,4-dione (6PPD-quinone) (2), hexa(methoxymethyl)melamine (HMMM) (3), and 1,3-diphenylguanidine (DPG) (4) have received increasing attention because of their environmental toxicity. However, the toxicological impact of urban runoff remains elusive because of the heterogeneous nature of runoff samples, which vary with weather conditions, and the large number of chemical constituents, which challenges data analysis.
Non-target screening (NTS) is essential to identify unknown micropollutants that may contribute substantially to environmental toxicity (5). Unlike targeted monitoring methods that focus on quantification of selected analytes using chemical reference standards, NTS aims to detect both known and unknown micropollutants to provide a more complete determination of the chemical composition (6). High-resolution mass spectrometry (HRMS) instruments enable this by generating data on molecular formulas and fragmentation patterns. To enhance identification of novel compounds, in-silico tools such as Sirius (7) can be used for annotation by calculating structural fingerprints based on isotopic and fragmentation data, enabling searches in structure databases such as PubChem (8) containing millions of chemical compounds.
Despite these identification tools, NTS of environmental samples remains challenging because of the complexity of the sample matrices. Recent studies have demonstrated that the number of acquired MS/MS spectra in data-dependent acquisition (DDA) drops sharply in complex environmental samples compared to clean matrices (9). Data-independent acquisition (DIA), which fragments all precursor ions, can address this but produces complex data sets that require peak deconvolution and generally results in more noisy MS/MS spectra. Another approach is to use iterative DDA injections that can improve MS/MS coverage (number of fragmented precursors) (10),though at the cost of longer analysis times and the generation of data less suited for quantification. A combined DDA-DIA strategy offers complementary benefits by providing full-scan fragmentation and quantitative data from DIA together with clean MS/MS spectra from DDA that can be used for annotation.
Here, we present a combined DDA-DIA workflow for chemical fingerprinting of urban runoff. The method uses iterative DDA on a pooled sample for structural annotation in Sirius and matches the resulting features to DIA data from individual samples. A schematic overview is shown in Figure 1. Using this approach, we tentatively identified several tire-related micropollutants, including two that have not been previously reported in environmental samples.
Materials and Methods
Sampling and Sample Preparation
Thirty-eight runoff samples were collected from different types of catchment areas in Santander (Spain), Pontedera (Italy), Ljubljana (Slovenia), Riga (Latvia), Odense (Denmark), and Copenhagen (Denmark) (see Table I for sample details). Field blank samples consisting of pure water with 300 mg/L CaCl2 were also collected for every 10 runoff samples. Samples were filtered through 0.7-µm glass fiber filters (Whatman) and extracted by multi-layer solid-phase extraction (ML-SPE) using 250-mg Supelclean envi-Carb columns (Merck) with 550 mg of a 1:1 mixture of Oasis HLB (Waters) and Isolute ENV+ sorbents (Biotage).
For DDA analysis, eight pooled samples were prepared by combining two to four individual extracts from similar catchment areas to represent different types of urban runoff. A pooled quality control sample (QCpooled) was also prepared by combining all extracts. QCpooled was injected every eight samples in the DIA sequence and analyzed in the DDA workflow like the other pooled samples.
Separation and Ionization Parameters
Analysis was performed using a Vanquish Ultra High Performance Liquid Chromatograph (UHPLC) (Thermo Scientific) coupled to an Orbitrap Exploris 240 MS (Thermo Scientific). Separation was achieved on a 100 mm × 2.1 mm, 1.7-µm BEH C18 column (Waters) using a gradient of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) at 0.3 mL/min. The gradient held at 1% B for 0.5 min, ramped to 20% at 2 min, 35% at 8 min, 75% at 18 min, and 99% at 21 min, held until 23.5 min, then returned to 1% B by 25 min.
Electrospray ionization (ESI) was used with sheath gas at 40 and aux gas at 5 (arbitrary units). The ion transfer tube was set to 325 °C and vaporizer temperature to 400 °C. Spray voltage was 4 kV (positive mode) and 2.5 kV (negative mode). Internal mass calibration was applied scan-to-scan for optimal mass accuracy.
Iterative DDA and DIA Analysis
DDA analysis was performed on pooled samples separately using Deep Scan in AcquireX (Thermo Fisher). For each of the samples, the composite field blank was first analyzed in full scan (resolving power 120,000) to generate an exclusion list, followed by a full scan of the sample to create an inclusion list. Samples were then run as three iterative DDA cycles, each including alternating full scans (60,000 resolving power) and four MS/MS experiments (30,000 resolving power) using stepped collision energies (20, 40, and 60 V). Exclusion lists were updated after each DDA run.
The DDA .raw files were converted to mzML and processed in Sirius (v5.8.6) for peak detection and annotation based on isotope patterns and fragmentation trees (7). Components were annotated by searching the full PubChem database. Annotation lists (mass-to-charge ratio [m/z] and retention times) for 4630 components in positive electrospray ionization (ESI+) and 1812 in negative electrospray ionization (ESI-) were exported as .txt files for matching to DIA features. Forty-four target compounds previously identified in raw data were used to validate annotation performance. Identifications were manually inspected, and confidence was scored according to the scheme shown in Figure 1.
DIA was performed using triplicate injections alternating between full scan (120,000 resolving power) and all-ion fragmentation (AIF) scans (30,000 resolving power) with stepped collision energies of 20 V, 40 V, and 60 V.Peak detection was done in MSDial version 4.92 (11)with 0.2 min retention time and 0.005 Da MS1 tolerance. For matching DIA features with annotated components from Sirius, the annotation .txt files were used for post-identification.
Drift correction was performed usingLOESS regression based on QCpooled injections every eight samples. Features were kept only if their average signal exceeded 50 times the average field blank signal and relative standard deviation (RSD) in triplicates was <50%. Additional filters removed features with >5 ppm m/z or >0.2 min retention time deviation between DDA and DIA data.
Results and Discussion
Evaluation of Component Matching from DDA to DIA
Out of 64,175 features detected in the DIA data, 4718 (6%) passed all filtering criteria and were matched to annotated DDA components. While few studies have assessed MS/MS coverage in complex environmental samples, this match rate is consistent with findings in recycled textiles (12) , where 3–5% of features were retained in ESI+ and ESI-. The filtering steps were designed to remove noise, and the resulting data reduction reflects a focus on components with high-quality MS/MS spectra and thus a higher likelihood of successful identification.
Figure 2 shows that matched features had significantly higher signal intensity (median: 7.9 × 104) than non-matched features (median: 2.8 × 104), indicating that the workflow effectively retained the more intense signals. Although high intensity does not necessarily correlate with environmental relevance, these features typically have cleaner MS/MS spectra, which improves annotation success.
Evaluation of Annotation Performance
Among 44 target compounds detected in the runoff samples, 22 (50%) were correctly identified as the top-ranked structure candidate by Sirius (Figure 3), while an additional eight (18%) were manually identified, either as lower-ranked candidates or by searching for the known m/z in the data.
This performance exceeds the 19–33% correct top-ranked annotations previously found in environmental water samples using single-injection DDA (9). This difference is likely due to our use of iterative injections in DDA, which increases the number of fragmented precursors, as demonstrated in metabolomics research (10). These findings demonstrate the advantage of iterative DDA for NTS, particularly when combined with pooled samples to minimize additional instrument time. In our study, incorporating iterative DDA analysis (in addition to triplicate analysis of individual samples and QC injection for every eight runs) increased the total number of injections by 35%, which is reasonable for studies aiming at the identification of unknown micropollutants.
Identification of Urban Runoff Pollutants
To assess pollution sources, we focused on components with high occurrence in street runoff samples. Several compounds were identified, as shown in Table II, with identification points from 100 for HMMM and DPG, which were confirmed with analytical standards to 45 for 2-cyclohexyl-1-phenylguanidine (CPG) and 2-butyl-1,3-bis(4-methylphenyl)guanidine (BBG), which have not previously been reported in environmental samples (additional compounds details in Table III) .
Compounds such as CPG and BBG, while unconfirmed, were frequently detected in street runoff samples (Figure 4) and may originate from tire wear, similar to the structurally related 1,3-di-o-tolylguanidine (DTG) and phenylguanidine (PHG), which have been reported in street runoff samples (13,14). With detection frequencies of 47–58% across runoff samples, CPG and BBG are therefore candidates for future monitoring of urban runoff.
Of the HMMM-related compounds, HMMM, diformylated HMMM (DiHMMM), and hexamethylolmelamine pentamethyl ether (HMPE) have previously been detected in traffic runoff (15). This corresponds well with our frequent detections in street runoff samples of all the HMMM-related compounds, as shown in Figure 4. TP215 has previously been identified as a HMMM transformation product present in wastewater and surface water (3). Vehicle traffic is therefore a likely source of this compound as well, which is supported by the correlation with HMMM in our study (Pearson correlation coefficient: 0.9).
Interestingly, TP215 showed the highest signal in an artificial football field (OD06) and was detected in all artificial football fields with rubber infill (CO09 had cork infill). This underlines the potential leaching from rubber infill used on football fields, which are typically made from repurposed tire materials. The high solubility of TP215 (logD = -1.5 at pH 7) may explain its widespread occurrence compared to less polar HMMM derivatives. It is therefore a relevant compound for future monitoring due to its high mobility in the environment and high detection frequency (84%) across runoff samples in our study.
Conclusion
We have developed a combined DIA and iterative DDA workflow for in-silico identification of unknown compounds in complex environmental samples. The approach improved annotation performance, with 50% of target compounds correctly identified as the top-ranked Sirius candidate and an additional 18% identified through manual inspection. Applied to urban runoff, we tentatively identified several tire-related compounds, including two that were not previously reported in environmental samples. The method provides an efficient workflow for prioritizing unknown contaminants. Future applications may integrate toxicity prediction tools such as MS2Tox(16) and MLinvitroTox (17) to support risk-based prioritization.
Acknowledgments
This study is a contribution to the European Union’s Horizon Europe project D4RUNOFF under grant agreement no. 101060638.
References
(1) Müller, A.; Österlund, H.; Marsalek, J.; Viklander, M. The Pollution Conveyed by Urban Runoff: A Review of Sources. Sci. Total Environ. 2020, 709, 136125. DOI: 10.1016/j.scitotenv.2019.136125
(2) Tian, Z.; Zhao, H.; Peter, K. T.; et al. A Ubiquitous Tire Rubber-Derived Chemical Induces Acute Mortality in Coho Salmon. Science 2021, 371, 185–189. DOI: 10.1126/science.abd6951
(3) Alhelou, R.; Seiwert, B.; Reemtsma, T. Hexamethoxymethylmelamine – A Precursor of Persistent and Mobile Contaminants in Municipal Wastewater and the Water Cycle. Water Res. 2019, 165, 114973. DOI: 10.1016/j.watres.2019.114973
(4) Zhang, H. Y.; Han, Y.; Hu, L. X.; et al. A Comprehensive Reconnaissance and Risk Assessment of Rubber Additives and Their Transformation Products (RATPs) in Groundwater: 1,3-Diphenylguanidine (DPG) as a Pressing Ecological Concern. Water Res. 2025, 277, 123279. DOI: 10.1016/j.watres.2025.123279
(5) Escher, B.; van Daele, C.; Dutt, M.; Tang, J.; Altenburger, R. Most Oxidative Stress Response In Water Samples Comes From Unknown Chemicals: The Need For Effect-Based Water Quality Trigger Values. Environ. Sci. Technol. 2013, 47, 7002–7011. DOI: 10.1021/es304793h
(6) Hollender, J.; Schymanski, E. L.; Ahrens, L.; et al. NORMAN Guidance on Suspect and Non-Target Screening in Environmental Monitoring. Environ. Sci. Eur. 2023, 35, 75. DOI: 10.1186/s12302-023-00779-4
(7) Dührkop, K.; Fleischauer, M.; Ludwig, M.; et al. SIRIUS 4: A Rapid Tool for Turning Tandem Mass Spectra into Metabolite Structure Information. Nat. Methods 2019, 16 (4), 299–302. DOI: 10.1038/s41592-019-0344-8
(8) Dührkop, K.; Shen, H.; Meusel, M.; Rousu, J.; Böcker, S. Searching Molecular Structure Databases with Tandem Mass Spectra Using CSI:FingerID. Proc. Natl. Acad. Sci. U S A 2015, 112 (41) 12580–12585. DOI: 10.1073/pnas.1509788112
(9) Peets, P.; Rian, M. B.; Martin, J. W.; Kruve, A. Evaluation of Nontargeted Mass Spectral Data Acquisition Strategies for Water Analysis and Toxicity-Based Feature Prioritization by MS2Tox. Environ. Sci. Technol. 2024, 58, 17406–17418. DOI: 10.1021/acs.est.4c02833
(10) Cooper, B.; Yang, R. An Assessment of AcquireX and Compound Discoverer Software 3.3 for Non-Targeted Metabolomics. Sci. Rep. 2024, 14, 4841. DOI: 10.1038/s41598-024-55356-3
(11) Tsugawa, H.; Cajka, T.; Kind, T.; et al. MS-DIAL: Data-Independent MS/MS Deconvolution for Comprehensive Metabolome Analysis. Nat. Methods 2015, 12 (6), 523–526. DOI: 10.1038/nmeth.3393
(12) Szabo, D.; Fischer, S.; Mathew, A. P.; Kruve, A. Prioritization, Identification, and Quantification of Emerging Contaminants in Recycled Textiles Using Non-Targeted and Suspect Screening Workflows by LC-ESI-HRMS. Anal. Chem. 2024, 96 (35), 14150–14159. DOI: 10.1021/acs.analchem.4c02041
(13) Gasperi, J.; Le Roux, J.; Deshayes, S.; et al. Micropollutants in Urban Runoff from Traffic Areas: Target and Non-Target Screening on Four Contrasted Sites. Water (Basel) 2022, 14 394. DOI: 10.3390/ w14030394
(14) Du, B.; Tian, Z.; Peter, K. T.; Kolodziej, E. P.; Wong, C. S. Developing Unique Nontarget High-Resolution Mass Spectrometry Signatures to Track Contaminant Sources in Urban Waters. Environ. Sci. Technol. Lett. 2020, 7, 923–930. DOI: 10.1021/acs.estlett.0c00749
(15) Peter, K. T.; Tian, Z.; Wu, C.; et al. Using High-Resolution Mass Spectrometry to Identify Organic Contaminants Linked to Urban Stormwater Mortality Syndrome in Coho Salmon. Environ. Sci. Technol. 2018, 52 (18), 10317–10327. DOI: 10.1021/acs.est.8b03287
(16) Peets, P.; Wang, W. C.; Macleod, M.; et al. MS2Tox Machine Learning Tool for Predicting the Ecotoxicity of Unidentified Chemicals in Water by Nontarget LC-HRMS. Environ. Sci. Technol. 2022, 56 (22), 15508–15517. DOI: 10.1021/acs.est.2c02536
(17) Arturi, K.; Harris, E. J.; Gasser, L.; et al. MLinvitroTox Reloaded for High-Throughput Hazard-Based Prioritization of High-Resolution Mass Spectrometry Data. J. Cheminform. 2025, 17 (1), 14. DOI: 10.1186/s13321-025-00950-4
Thomas Molnár Karlsson is a PhD student in analytical chemistry at the Department of Plant and Environmental Sciences at the University of Copenhagen. His research focuses on novel detection methods for non-target screening of urban runoff pollutants and is funded by the Horizon Europe 2021 project D4Runoff.
Jan H. Christensen is a professor in environmental analytical chemistry. He is leader of the Analytical Chemistry group, Faculty of Science, University of Copenhagen, and heads the Research Centre for Advanced Analytical Chemistry (RAACE).
![Figure 4: Heatmap of identified compounds from different runoff catchment areas (streets, combined sewer overflows [CSO], roofs, residential areas, and artificial football fields), showing detection frequency and relative signal intensity (range scaled 0–100). Gray boxes indicate non-detections. Compounds include confirmed and tentatively identified tire-related pollutants, with high occurrence observed in street runoff samples.](/_next/image?url=https%3A%2F%2Fcdn.sanity.io%2Fimages%2F0vv8moc6%2Fchroma%2Fdb4d9e2dd0c93c8e93a37a9178c8ea24ed5066e7-652x238.jpg%3Ffit%3Dcrop%26auto%3Dformat&w=1920&q=75 "Figure 4: Heatmap of identified compounds from different runoff catchment areas (streets, combined sewer overflows [CSO], roofs, residential areas, and artificial football fields), showing detection frequency and relative signal intensity (range scaled 0–100). Gray boxes indicate non-detections. Compounds include confirmed and tentatively identified tire-related pollutants, with high occurrence observed in street runoff samples.")
