Earth Day 2026: Assessing Marine Pollution and Ecosystem Impacts Through LC-HR-MS/MS Analysis of Anthropogenic Compounds in Oceanic Dissolved Organic Matter

Intensifying population growth, urbanization, industrialization, and consumer-driven activities have led to a surge of chemical pollutants entering marine environments, threatening biodiversity, disrupting food webs, and degrading essential ecosystem services. Fortunately, combining test results from multiple laboratories and studies provides an exciting new way to track human-made chemicals and study organic matter across the globe.¹

A research team recently applied non-targeted liquid chromatography–high-resolution tandem mass spectrometry (LC-HR-MS/MS) to investigate the extent of anthropogenic organic contaminants within the dissolved organic matter (DOM) pool. By integrating 21 publicly available datasets spanning over 2300 samples from diverse marine regions, the research overcomes the spatial limitations of individual LC-MS studies. Using standardized chromatographic methods, the analysis reveals that xenobiotic compounds are not only widespread across coastal, coral reef, and open ocean systems, but also form a significant fraction of marine DOM. These findings underscore the growing environmental burden of chemical pollution in oceans and highlight the need for large-scale monitoring strategies to better understand and mitigate their ecological and biogeochemical impacts.

LCGC International spoke to Daniel Petras and Jarmo Kalinski, two of the authors of the resulting paper published in Nature Geoscience,² about the work.

How does liquid chromatography (LC) enable the separation of complex dissolved organic matter (DOM) mixtures before MS detection?
Daniel Petras: For us, liquid chromatography is essential to separate isobars and isomers before MS/MS analysis as much as possible. Without chromatography, most MS/MS spectra we acquire would be mixtures of dozens, if not hundreds, of different precursors that fall within the quadrupole isolation window. In fact, even with UHPLC, we still observe a large number of mixed (chimeric) spectra, which is why we are very interested in improving our chromatographic methods, for example through 2D-LC, or by adding orthogonal separation methods such as ion mobility.

What are the advantages of coupling liquid chromatography with high-resolution tandem mass spectrometry (LC-HR-MS/MS) for xenobiotic analysis in marine systems?
Jarmo Kalinski: Coupling liquid chromatography with high-resolution tandem mass spectrometry gives us a powerful window into xenobiotics within the marine dissolved organic matter pool. By separating compounds chromatographically before they enter the mass spectrometer, we can detect tens of thousands of features in a single run while maintaining high sensitivity, often down to the picogram-per-liter range in seawater.

The high mass accuracy provided by HRMS allows us to determine elemental compositions. When we combine that with MS/MS fragmentation spectra, we gain structural information that helps us putatively identify compounds, even when analytical standards are not available. Tools such as spectral-library matching and MS/MS-based in silico structure prediction become especially valuable in this context.

This is critically important in marine DOM, where xenobiotic signatures are typically subtle, highly diluted, and embedded within a massive natural organic background. LC-HR-MS/MS gives us the resolution, accuracy, and sensitivity needed to distinguish these anthropogenic compounds from the natural DOM matrix and to begin assigning meaningful structures to them.

What chromatographic parameters (e.g., mobile phase, column type, gradient) would you optimize when analyzing marine DOM samples?

DP: We very much like C18 reversed-phase columns and are currently using a core-shell column with organo-silica bridged 1.7 µm particles and dimensions of 150 × 2 mm for all our DOM analyses. A nice feature of the hybrid silica material is that we can also run these columns at high pH, which is ideal for our high/low-pH 2D-LC. Our gradients are typically 10 minutes long and run from 5 to 99% acetonitrile, with formic acid as the default modifier.

How do matrix effects from seawater (e.g., salts, natural organic matter) impact chromatographic separation and detection, and how can they be mitigated?
JK: Matrix effects can lead to subtle shifts in retention times between runs, and high salt content suppresses ionization efficiency, thereby lowering sensitivity, while also promoting adduct formation and altering mass spectrometric patterns. Similarly, the natural organic matter background affects ionization efficiency and can suppress, or even enhance, the ionization of xenobiotic signals. That is why we typically perform SPE before LC-MS/MS analysis. To mitigate retention time shifts, especially over large sample batches or between datasets, we employ wider retention time tolerances during the alignment step in data processing or use retention time correction in mzmine.

Why is non-targeted LC-HR-MS/MS particularly suitable for detecting unknown xenobiotics in marine environments?
DP: For us, non-targeted LC-HR-MS/MS is particularly well suited for detecting unknown xenobiotics in marine environments because it covers a broad range of chemical space, including a wide polarity and size range, and does not require us to know in advance which compounds we are looking for. Marine dissolved organic matter is chemically extremely complex, and many anthropogenic contaminants are unexpected, transformed, or absent from standard target or suspect lists.

What challenges arise when comparing chromatographic data across multiple laboratories, even with standardized methods?
JK: Two main challenges arise. First, retention time shifts caused by different LC columns, differences in delay volume or other system-specific parameters can create problems when aligning data from different laboratories. This necessitates the use of wider retention time alignment windows, which in turn increases the risk of false alignments. Another challenge is differences in the quality of MS/MS spectra between instruments, as well as the presence of chimeric MS/MS spectra in highly complex samples, meaning MS/MS spectra resulting from coeluting features with m/z values within the quadrupole isolation window. This can make it difficult to align chromatographic runs from different laboratories based solely on MS/MS spectral similarity.

One solution is to include shared standard mixtures containing compounds that span a wide polarity range, and therefore elute across the chromatographic run, as QC samples in experiments to be compared. These could be used for improved retention time alignment, similar to how n-alkane mixes are used to determine Kovats indices in GC. For MS/MS-based comparisons across laboratories, dedicated workflows that leverage retention time alignment in addition to MS/MS similarity are especially helpful.

How does retention time alignment contribute to meta-analysis of LC-MS datasets from different studies?
JK: Retention time alignment is one way for the integration of mass spectrometric signals across laboratories, especially if MS/MS quality is not consisten (e.g. due to chimeric spectra). In our workflow MS1 alignment is the first step for cross-study comparisons. In addition we combine the MS1-based alignment with MS/MS based annotation and merging of redundant features, which improves confidence in compound reoccurrence across datasets.

What role does chromatographic separation play in improving structural elucidation of compounds in tandem MS (MS/MS)?
DP: I think good chromatography is extremely important for obtaining high-quality MS/MS data, especially from ultracomplex samples such as DOM. Having sharper peaks and better separation ultimately results in better signal-to-noise ratios and cleaner MS/MS spectra, which is essential for high-quality spectrum-library matching and especially for in silico annotation.

I also feel that, particularly in a world of ultrahigh-resolution MS, chromatography is sometimes a little underappreciated, especially since the mass spectrometer is the biggest cost. In our recent interlaboratory comparison of DOM analysis across more than 20 labs, we got a good idea of which parameters are key to obtaining the most in-depth results from DOM samples. I can tell you that using a 5 µm particle-size column instead of a 1.7 µm particle-size UHPLC column makes a big difference in peak shape and overall annotation results.

Why is LC-HR-MS/MS typically limited to small spatial scales, and how can meta-analysis overcome this limitation?
JK: Research groups generally optimize their chromatographic and mass spectrometric parameters for the specific research question at hand, and this leads to significant variability between laboratories, even when the same samples are analyzed. Moreover, the logistical challenges of performing large-scale analyses within a single laboratory are substantial. Meta-analysis can collate multiple studies, but the overall confidence in the results will naturally be somewhat lower than for data recorded in a single experiment. The use of shared QC samples across laboratories and dedicated workflows for interlaboratory alignment and normalization are powerful approaches to address these limitations.

How can chromatographic reproducibility influence the reliability of large-scale environmental conclusions about xenobiotic distribution?
DP: Chromatographic reproducibility is a key factor for meta-analysis, especially for complex samples and MS1-based alignment. Ideally, we would all run our samples using standardized methods across laboratories, with community-based QC samples to ensure the best comparability. The reality, of course, is very different, with many different methods and systems being used across laboratories. Even in our own case, where we co-analyzed samples from only three laboratories using highly standardized methods, we could clearly see batch effects.

In this context, it is important to use data-analysis strategies that enable effective alignment. In our case, we used wider tolerances during chromatographic alignment and merged features based on MS/MS level 2 IDs, which is an effective way to bridge redundant features across datasets. Therefore, the effect of misalignment on our conclusions about xenobiotic distribution should be minimal, because a given annotated compound, for example cocaine in sample A, will also be annotated as cocaine in sample B, whether the chromatographic peaks match perfectly or not. For unannotated features, this is of course more difficult, but we would not consider those in our xenobiotic analysis anyway, since we do not know their identity or origin.

Looking forward, different software packages such as mzmine include modules for retention time correction, and MS2-based alignment approaches such as classical molecular networking bypass retention time entirely. My guess is that in the future we will use some hybrid approach that leverages both retention time normalization or correction and MS2-based alignment.

References

1. Kalinski, J. C.; Ruiz Brandão da Costa, B.; Schramm, T. et al. Comparability of Liquid Chromatography Tandem Mass Spectrometry Analysis of Dissolved Organic Matter across Laboratories. Environ Sci Technol. 2026, 60 (6), 4814-4829. DOI: 10.1021/acs.est.5c12691
2. Kalinski, J. C. .J.; Pakkir Mohamed Shah, A. K.; Ruiz Brandão da Costa, B. et al. Widespread Presence of Anthropogenic Compounds in Marine Dissolved Organic Matter. Nat. Geosci. 2026. DOI: 10.1038/s41561-026-01928-z