MSACL 2026 Abstract

Rapid Chromatography-Free Mass Spectrometry for High-Throughput Metabolomics and Lipidomics Across Diverse Biological Matrices

Felix Schmauder (1), Daniel Simon (5), Xiao Wu (2), Adrien Bitton (3), Pierre-Maxence Vaysse (4), Amalthé Paupiere (6), Paul Duhamel (6), Aurore Tomezyk (6), Guillaume Hochart (6), Mathieu Gaudin (6), István Pap (5), Gabriel Horkovics-Kovats (5), Zóltan Takáts (5), Oleksandr Boychenko (1)
(1) Bayer AG, Pharmaceuticals – Translational Sciences, Wuppertal, Germany (2) Bayer AG, Consumer Health – Steigerwald Arzneimittel, Darmstadt, Germany (3) Bayer AG, Pharmaceuticals – DS&AI, Berlin, Germany (4) Bayer SAS, Crop Science – Human Safety, Sophia Antipolis, France (5) University of Regensburg, Regensburg, Germany (6) Aliri Bioanalysis – Spatial Bioanalysis Team, Lille, France

Felix Schmauder (Presenter) Bayer AG

Presenter Bio: Since 2025: Post-Doc at Bayer AG, Wuppertal
2021-2025: PhD-Studies, Hamburg School of Food Science, University of Hamburg, Germany
2015-2021: Studies of Foodchemistry, University of Würzburg, Germany

Relevant Financial Disclosures (within past 24 months, reported on Apr 22, 2026) Salary Bayer AG Discussing Products? NO, presenter will NOT present educational content related to products/services of the ineligible company(ies) listed above. Conflict mitigation NOT REQUIRED.

INTRODUCTION:
Conventional LC-MS metabolomics provides deep biochemical insights but is limited by long chromatographic run times, restricting throughput and number of analyzed samples in reasonable time. Chromatography-free MS approaches such as MALDI-MS, Direct Infusion-MS and Laser-REIMS offer rapid, within minutes, acquisition of metabolomic and lipidomic profiles while maintaining biologically meaningful depth.

OBJECTIVES:
This work aims to evaluate and optimize chromatography‑free MS workflows for high‑throughput profiling of plasma and urine from biobanks and large clinical study cohorts. The resulting data enables mechanistic interpretation as well as machine‑learning‑supported subjects stratification and phenotyping.

METHODS:
We established a semi-automated sample preparation to deliver samples for multi-platform data generation using MALDI-MS (Bruker SolariX), DI-MS (Thermo Fisher, Astral) and Laser-REIMS (REIMS-Interface, MediMass; Xevo G2-XS qToF Waters Corporation). Methods were optimized for speed, reproducibility and metabolome/lipidome coverage across matrices. DI-MS, MALDI-MS and Laser-REIMS offered the possibility to generate above 500 fingerprints per day. Unified data processing pipelines, classification using random forest and clustering models were implemented to evaluate metabolic differences and identify discriminatory molecular markers between cohorts and disease indications. 

RESULTS:
Chromatography-free MS enabled robust acquisition of metabolomic and lipidomic fingerprints from 6500 samples. Direct-MS methods delivered deep profiling compatible with studies involving thousands of samples. MS fingerprints retained biological information sufficient for mechanistic interpretation and ML-based molecular marker discovery. Machine-learning models identified metabolites and lipids differentiating biological states in disease-, treatment- and recovery- and allowed to link these to biological pathways.

CONCLUSION:
Chromatography-free MS technologies enable rapid, scalable metabolomics and lipidomics suitable for high-throughput data generation enabling the creation and utilization of ML-models. These platforms reduce measurement times and costs by an order of magnitude and deliver actionable metabolic signatures that support mechanistic insights, biomarker discovery, and data‑driven decision‑making.