A Novel Two-Step Workflow for Extracting Clean Mass Spectra in LC×LC–HRMS Data

LCGC International spoke to Paul-Albert Schneide and Oskar Munk Kronik from the Department of Food Science and the Department of Plant and Environmental Sciences at the University of Copenhagen (Copenhagen, Denmark) about the development and application of a novel two-step workflow—mass filtering (MF) combined with multivariate curve resolution (MCR)—for extracting clean mass spectra from trace-level compounds in comprehensive two-dimensional liquid chromatography high-resolution mass spectrometry (LC×LC–HRMS) data. They discuss the complexities of data processing in two-dimensional liquid chromatography, the advantages of MCR in deconvoluting data-independent acquisition (DIA) spectra, and the superior performance of MF + MCR compared to conventional extraction methods.

What is the main goal of the two-step workflow, MF + MCR, developed and discussed in your paper (1)?

Schneide: The main goal of the two-step workflow is to facilitate the extraction of clean mass spectra from compounds at trace-level, in applications where a list of known suspects is available. Mass filtering (MF) removes most mass fragments that have low similarity with the EIC of the target mass, while multivariate curve resolution (MCR) is very useful for the final cleaning step by removing chemical and electronic noise as well as signals from co-eluting compounds. Brought together, the workflow helps to improve the certainty in compound annotations based on fragment mass spectra.

Why is data processing in LC×LC–HRMS more complex than in one-dimensional liquid chromatography?

Schneide: One big challenge comes with the data size itself, which makes manual data analysis unfeasible. Furthermore, comprehensive two-dimensional liquid chromatography high-resolution mass spectrometry (LC×LC–HRMS) is still a relatively young analytical technique, and the focus of the community has been more on hardware innovation rather than developing data analysis capabilities. More mature data analysis workflows exist for LC–HRMS/MS. However, conceptually, the second dimension enables the use of algorithmic approaches for data analysis that are less effective or not possible to use for one-dimensional data. For example, the same mass-filtering approach that we have proposed would be less effective on one-dimensional data because the second dimension provides additional selectivity. Moreover, multi-linear decomposition methods can exploit the second-order advantage that LC×LC–HRMS intrinsically offers.

What challenge in data-independent acquisition (DIA) mode does MCR help to address?

Munk Kronik: In DIA mode, all precursor ions in a mass spectrum are fragmented simultaneously, producing a fragment mass spectrum that may contain a mixture of fragments, from chromatographically co-eluting compounds to baseline compounds. Curve-resolution techniques, such as MCR, can be used to mathematically deconvolute these spectra into cleaner mass spectra, ideally representing individual compounds.

What sample was used to test the effectiveness of the MF + MCR workflow, and why was it chosen?

Munk Kronik: We tested the effectiveness of the MF + MCR workflow using a pulsed elution LC×LC–HRMS chromatogram of a wastewater sample, chosen for its high chemical complexity. This chromatogram was published in a previous paper, where the focus was on optimizing the method parameters of pulsed elution LC×LC–HRMS (2).

How did the performance of MF + MCR compare to other mass spectra extraction methods in terms of compound identification?

Schneide: We could show on spiked standards that the MF + MCR workflow performs significantly better than a workflow based on spectra extraction at peak apex and significantly better than MCR without further pre-processing. Also, in the wastewater sample that we investigated, we could confirm the presence of 25 suspects with the MF+MCR approach, while only 10 and 16 suspects could be confirmed using the peak apex and the MCR approach, respectively.

Why is preprocessing before MCR important for improving compound detection?

Schneide: This is a bit nerdy, and it has to do with the way MCR solutions are calculated using alternating least squares. These algorithms usually assume a so-called “low rank” structure of the data, which means that a small region of the LC×LC–HRMS data set can be approximated well by only a few “latent variables”. However, this concept doesn’t work for trace compounds where the analyte signal only contributes marginally to the variation in the data.

Rob Synovec and his group have described similar problems when using MCR (or PARAFAC) for the analysis of trace-level compounds in comprehensive two-dimensional gas chromatography (GC×GC) combined with time-of-flight mass spectrometry (TOF-MS) (3). In the context of hyperspectral imaging, alternative approaches to address this problem have been developed, which I assume are worth exploring for applications to two-dimensional chromatography (4,5).

How can the MF + MCR workflow benefit scientists working in environmental analysis? Can this workflow be used in any other application areas?

Munk Kronik: Scientists in environmental analysis can benefit from the MF + MCR workflow for more reliable compound identification because of the higher quality of the mass spectra obtained. In this study, we demonstrated that low-intensity compounds could be reliably identified using the MF + MCR workflow. This approach is particularly relevant for environmental monitoring of chemicals of emerging concern, even at trace levels.

We believe the workflow is useful in other applications as well. People working in applications with a high number of isomeric compounds, such as in plant analysis, should be cautious not to exclude relevant ions during the MF step.

References

(1) Schneide, P.-A.; Kronik, O. M. Signal Processing Workflow for Suspect Screening in LC × LC-HRMS: Efficient Extraction of Pure Mass Spectra for Identification of Suspects in Complex Samples Using a Mass Filtering Algorithm. Anal. Chem. 2025, 97 (2), 1180–1189. DOI: 10.1021/acs.analchem.4c04288

(2) Kronik, O. M.; Christensen, J. H.; Nielsen, N. J. Instrumental and Theoretical Advancements in Pulsed Elution-LC × LC: Investigation of Pulse Parameters and Application to Wastewater Effluent. J. Chrom A 2024, 1730, 465079. DOI: 10.1016/j.chroma.2024.465079

(3) Ochoa G.S.; Sudol, P. E.; Trinklein, T. J.; Synovec, R. E. Class Comparison Enabled Mass Spectrum Purification for Comprehensive Two-dimensional Gas Chromatography with Time-of-Flight Mass Spectrometry. Talanta 2022, 236, 122844. DOI: 10.1016/j.talanta.2021.122844

(4) Coic, L.; Sacré, P.-Y.; Dispas, A.; et al. Selection of Essential Spectra to Improve the Multivariate Curve Resolution of Minor Compounds in Complex Pharmaceutical Formulations. Anal. Chim. Acta2022, 1198, 339532. DOI: 10.1016/j.aca.2022.339532

(5) Gallagher, N. B.; Goyetche, R.; Amigo, J. M.; Kucheryavskiy, S. Extended Least Squares (ELS) and Generalized Least Squares (GLS) for Clutter Suppression in Hyperspectral Images: A Theoretical Description. Chemom. Intell. Lab. Syst. 2024, 244, 105032. DOI: 10.1016/j.chemolab.2023.105032

Images courtesy of interviewees

Oskar Munk Kronik specializes in method development, instrument configurations, and data processing workflows for comprehensive 2D-LC. His research focuses on the role of 2D-LC in non-target screening. He is currently a Ph.D. stipend at the University of Copenhagen.

Paul-Albert Schneide’s research focuses on the development of chemometric and signal processing methods for complex instrumental analytical data, with a strong emphasis on hyphenated chromatographic techniques. He is currently employed as a postdoctoral researcher at the University of Copenhagen and as a research scientist in the Analytical Science Department of BASF SE.