EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Generalizable machine learning models for rapid antimicrobial resistance prediction in unseen health care settings

Diane Duroux iD, Paul P Meyer, Giovanni Visonà, Niko Beerenwinkel iD

GigaScience · 21.01.2026 · 15

DOI: 10.1093/gigascience/giaf156 ↗

Kurz gefasst

Fachartikel in der Zeitschrift GigaScience, online erschienen am 19.01.2026. Verfasst von Diane Duroux und 3 weiteren Autor:innen. Beteiligte Einrichtung: ETH AI Center, ETH Zurich, Andreasstrasse 5, 8092, Zurich, Switzerland. Hauptthemen laut MEDLINE: Machine Learning, Drug Resistance, Bacterial, Anti-Bacterial Agents. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: ETH.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The deployment of machine learning in clinical settings is often hindered by the limited generalizability of the models. Models that perform well during development tend to underperform in new environments, limiting their clinical utility. This issue affects models designed for the rapid identification of antimicrobial resistance, which is essential to guide treatment decisions. Traditional susceptibility tests can take up to 3 days, whereas integrating matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry with machine learning has the potential to reduce this to 1 day. However, model performance declines drastically in hospitals or time frames outside the training data.

Results

To improve robustness, we develop advanced feature representations using masked autoencoders (MAEs) for MALDI-TOF spectra and chemical language models and SELF-referencing embedded strings (SELFIES) for antimicrobials. Cross-validated on data from 4 medical institutions, our models demonstrate improved performance and stability. The MAE and SELFIES encodings increase the area under the precision–recall curve by 4% when evaluated on unseen time periods, while the MAE and Molformer language model encodings improve it by 10% when applied across different hospitals.

Conclusions

These results underscore the value of combining deep learning with chemical and spectral information to build generalizable, high-impact clinical artificial intelligence.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningDrug Resistance, BacterialAnti-Bacterial AgentsPredictive Learning ModelsAutoencoderHumansSpectrometry, Mass, Matrix-Assisted Laser Desorption-IonizationMicrobial Sensitivity Tests

Stichwörter der Autor:innen

antimicrobial resistancegeneralizabilitymachine learningrepresentation learning

Substanzen

Anti-Bacterial Agents

Autor:innen und Institutionen

  • ETH AI Center, ETH Zurich, Andreasstrasse 5, 8092, Zurich, Switzerland.
  • Department of Biosystems Science and Engineering, ETH Zurich, Klingelbergstrasse 48, 4056, Basel, Switzerland.
  • Department of Quantitative Biomedicine, University of Zurich, Schmelzbergstrasse 26, 8006, Zurich, Switzerland.
  • SIB Swiss Institute of Bioinformatics, Quartier UNIL-Sorge, 1015, Lausanne, Switzerland.
  • Department of Mechanical and Process Engineering, ETH Zurich, Tannenstrasse 3, 8092, Zurich, Switzerland.
  • Department of Empirical Inference, Max Planck Institute for Intelligent Systems, Max-Planck-Ring 4, 72076, Tübingen, Germany.

ORCID-Profile: Diane Duroux, Niko Beerenwinkel

Förderung

  • ETH

Zitieren

APA

Duroux, D., Meyer, P. P., Visonà, G., & Beerenwinkel, N. (2026). Generalizable machine learning models for rapid antimicrobial resistance prediction in unseen health care settings. GigaScience, 15. https://doi.org/10.1093/gigascience/giaf156

Vancouver

Duroux D, Meyer PP, Visonà G, Beerenwinkel N. Generalizable machine learning models for rapid antimicrobial resistance prediction in unseen health care settings. GigaScience. 2026;15. doi:10.1093/gigascience/giaf156

BibTeX
@article{Duroux2026,
  author = {Duroux, Diane and Meyer, Paul P and Visonà, Giovanni and Beerenwinkel, Niko},
  title = {Generalizable machine learning models for rapid antimicrobial resistance prediction in unseen health care settings},
  journal = {GigaScience},
  year = {2026},
  volume = {15},
  doi = {10.1093/gigascience/giaf156}
}
Kostenlos starten →