EUVIMEDEUVIMED entdecken
Crossref

Protein and genomic language models uncover the unexplored diversity of bacterial immunity

Ernest Mordret iD, Alexandre Hervé iD, Florian Tesson iD, Hugo Vaysset iD, Tyler Clabby iD, Arthur Loubat iD, Helena Shomar iD, Remi Planel iD, Rachel Lavenir iD, Jean Cury iD, Aude Bernheim iD

Science · 02.04.2026 · 392 (6793)

DOI: 10.1126/science.adv8275 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science, erschienen am 02.04.2026. Verfasst von Ernest Mordret und 10 weiteren Autor:innen. Beteiligte Einrichtung: Institut Pasteur, Université Paris-Cité, CNRS UMR3525, Molecular Diversity of Microbes, Paris, France. Hauptthemen laut MEDLINE: Bacterial Proteins, Bacteriophages, Escherichia coli, Genome, Bacterial. Die Arbeit wurde bisher 30-mal zitiert (Crossref) und nennt 62 Literaturangaben. Gefördert von: MSDAVENIR, European Research Council.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The bacterial pangenome contains a vast diversity of antiphage systems, whose overall extent is still unknown. In this study, we developed complementary machine learning approaches to systematically predict antiphage function from genomic context, protein sequence, or their combination, achieving up to 99% precision and 92% recall. We validated these models experimentally in Escherichia and Streptomyces with the discovery of 12 antiphage systems. Applied to over 32,000 bacterial genomes, these models expand the predicted antiphage repertoire, with ~1.5% of bacterial genomes devoted to defense and more than 85% of predicted protein families remaining uncharacterized. We provide an interactive catalog of more than 19,000 candidate operon families for experimental follow-up. Together, these findings show that most molecular diversity in bacterial immunity remains uncharacterized and provide a foundation for its systematic exploration.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Bacterial ProteinsBacteriophagesEscherichia coliGenome, BacterialMachine LearningStreptomycesGenomicsOperonLarge Language Models

Substanzen

Bacterial Proteins

Autor:innen und Institutionen

  • Institut Pasteur, Université Paris-Cité, CNRS UMR3525, Molecular Diversity of Microbes, Paris, France.
  • INSERM U1284, Université Paris-Cité, ELIS, Paris France.
  • AgroParisTech, Université Paris-Saclay, Paris, France.
  • Institut Pasteur, Université Paris Cité, Bioinformatics and Biostatistics Hub, Paris, France.

ORCID-Profile: Ernest Mordret, Alexandre Hervé, Florian Tesson, Hugo Vaysset, Tyler Clabby, Arthur Loubat, Helena Shomar, Remi Planel, Rachel Lavenir, Jean Cury, Aude Bernheim

Förderung

  • MSDAVENIR
  • European Research Council

Zitieren

APA

Mordret, E., Hervé, A., Tesson, F., Vaysset, H., Clabby, T., Loubat, A., Shomar, H., Planel, R., Lavenir, R., Cury, J., & Bernheim, A. (2026). Protein and genomic language models uncover the unexplored diversity of bacterial immunity. Science, 392(6793). https://doi.org/10.1126/science.adv8275

Vancouver

Mordret E, Hervé A, Tesson F, Vaysset H, Clabby T, Loubat A, et al. Protein and genomic language models uncover the unexplored diversity of bacterial immunity. Science. 2026;392(6793). doi:10.1126/science.adv8275

BibTeX
@article{Mordret2026,
  author = {Mordret, Ernest and Hervé, Alexandre and Tesson, Florian and Vaysset, Hugo and Clabby, Tyler and Loubat, Arthur and Shomar, Helena and Planel, Remi and Lavenir, Rachel and Cury, Jean and Bernheim, Aude},
  title = {Protein and genomic language models uncover the unexplored diversity of bacterial immunity},
  journal = {Science},
  year = {2026},
  volume = {392},
  number = {6793},
  doi = {10.1126/science.adv8275}
}
Kostenlos starten →