SP 1.2.1 - Machine Learning for longitudinal data analysis in gas sensors: A case study in early subclinical ketosis in dairy cows

Event

EUROSENSORS 2026
2026-09-06 - 2026-09-09
Zurich

Band

Lectures

Chapter

Antimicrobial Resistance Diagnostics

Author(s)

A. Belenguer-Llorens, C. R. Jutzeler - ETH Zurich,Zurich (Switzerland) and SIB Swiss Institute of Bioinformatics,Lausanne (Switzerland), J. Hendricks, A. -M. Reiche, F. Dohme-Meier - Agroscope,Posieux (Switzerland), J. Ko, A. T. Güntner - ETH Zurich,Zurich (Switzerland)

Pages

283 - 284

DOI

10.5162/eurosensors2026/SP1.2.1

ISBN

978-3-910600-12-6

Price

free

Abstract

Urinary tract infections caused by Escherichia coli are increasingly resistant to first-line antibiotics, yet current culture-based diagnostics require 2 to 4 days, driving empiric broad-spectrum therapy that accelerates resistance spread. Diagnostic gaps persist, with low-level and fastidious organisms frequently missed by standard methods, while rapid susceptibility testing remains unavailable in most settings. To address this challenge, we developed an untargeted and targeted metabolomics approach to identify urinary metabolite signatures specific to UTI-causing bacteria. Using 75 E. coli strains grown in artificial urine medium and analysed by complementary liquid chromatography–mass spectrometry methods, we identified differential metabolite production patterns including increased succinate and depleted mannitol in bacterial cultures. These metabolomic signatures offer potential biomarkers for rapid bacterial identification independent of culture. This approach could enable faster, more specific UTI diagnostics to reduce inappropriate antibiotic use and combat antimicrobial resistance.