EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Harnessing interpretable deep learning to predict resistance in Klebsiella pneumoniae

Nicolas da Matta Freire Araujo, Márcia da Silva Chagas, Mateus Fernandes Santos, Renata Freire Alves Pereira, Rafaela Correia Brum, Felipe Ramos Pinheiro, Melise Chaves Silveira, Felicita Mabel Duré, Beatriz de Lima Alessio Müller, Audrien Alves Andrade de Souza, Alessandra Beatriz Santos Rondon Souza, Ágatha Ferreira de Souza und 8 weitere

Frontiers in Cellular and Infection Microbiology · 2026

DOI: 10.3389/fcimb.2026.1859508 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Cellular and Infection Microbiology, 2026. Verfasst von Nicolas da Matta Freire Araujo und 19 weiteren Autor:innen. Beteiligte Einrichtung: Scientific Computing Program, Oswaldo Cruz Foundation (FIOCRUZ), Rio de Janeiro, Brazil. Hauptthemen laut MEDLINE: Klebsiella pneumoniae, Deep Learning, Anti-Bacterial Agents, Drug Resistance, Bacterial. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 73 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundação Oswaldo Cruz, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Antimicrobial resistance poses a growing global health threat, complicating therapeutic management and increasing morbidity and mortality. Deep learning methods have emerged as effective tools for bacterial profiling using omics data, particularly in predicting antimicrobial susceptibility from genomic information. The present study focuses on identifying genomic signatures linked to resistance mechanisms using a deep learning architecture.

Methods

DeepMDC, a deep learning architecture for bacterial profiling using whole-genome data, is introduced. Due to the high cost and ambiguity of precise gene- or mutation-level annotation, phenotypic classification is formulated as a Multiple Instance Learning (MIL) problem, in which each genome is represented as a bag of instances with a single associated label. The architecture centers on a modern Hopfield network that processes all open reading frames (ORFs), including small ORFs, derived from genomic data. Interpretability is achieved through attention mechanisms, which facilitate biological insight and support hypothesis generation.

Results

The model was evaluated against Klebsiella pneumoniae and four clinically relevant antibiotics (meropenem, cefepime, ceftazidime, and gentamicin), achieving strong performance across multiple evaluation metrics.

Discussion

Notably, genes associated with resistance consistently received high attention scores during inference, validating the architecture and potentially generating new hypotheses.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Klebsiella pneumoniaeDeep LearningAnti-Bacterial AgentsDrug Resistance, BacterialGenome, BacterialMicrobial Sensitivity TestsGenomicsOpen Reading FramesHumansKlebsiella InfectionsDrug Resistance, Multiple, Bacterial

Stichwörter der Autor:innen

Klebsiella pneumoniaeantimicrobial resistancedeep learninggenomicsmodern Hopfield networks

Substanzen

Anti-Bacterial Agents

Autor:innen und Institutionen

  • Scientific Computing Program, Oswaldo Cruz Foundation (FIOCRUZ), Rio de Janeiro, Brazil.
  • Department of Applied Mathematics, Institute of Mathematics, Federal University of Rio de Janeiro (UFRJ), Rio de Janeiro, Brazil.
  • Laboratory of Molecular Epidemiology and Biotechnology, School of Pharmacy/Fluminense Federal University, Niteroi, RJ, Brazil.
  • Graduate Program in Applied Microbiology and Parasitology. Biomedical Institute - Fluminense Federal University, Niteroi, RJ, Brazil.
  • Department of System Engineering and Computation, State University of Rio de Janeiro, Rio de Janeiro, Brazil.
  • Graduate Program in Pathology, Fluminense Federal University, Niteroi, RJ, Brazil.

Förderung

  • Fundação Oswaldo Cruz
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico
  • Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro

Zitieren

APA

Araujo, N. D. M. F., Chagas, M. D. S., Santos, M. F., Pereira, R. F. A., Brum, R. C., Pinheiro, F. R., Silveira, M. C., Duré, F. M., Müller, B. L. A., de Souza, A. A. A., Souza, A. B. S. R., de Souza, Á. F., Carvalho-Assef, A. P. D., Moreira, A. D. S., Dos Santos, M. T., Côrtes, A. M. A., Penna, B. A., Chagas, T. P. G., Aguiar-Alves, F., & da Silva, F. A. B. (2026). Harnessing interpretable deep learning to predict resistance in Klebsiella pneumoniae. Frontiers in Cellular and Infection Microbiology. https://doi.org/10.3389/fcimb.2026.1859508

Vancouver

Araujo NDMF, Chagas MDS, Santos MF, Pereira RFA, Brum RC, Pinheiro FR, et al. Harnessing interpretable deep learning to predict resistance in Klebsiella pneumoniae. Frontiers in Cellular and Infection Microbiology. 2026. doi:10.3389/fcimb.2026.1859508

BibTeX
@article{Araujo2026,
  author = {Araujo, Nicolas da Matta Freire and Chagas, Márcia da Silva and Santos, Mateus Fernandes and Pereira, Renata Freire Alves and Brum, Rafaela Correia and Pinheiro, Felipe Ramos and Silveira, Melise Chaves and Duré, Felicita Mabel and Müller, Beatriz de Lima Alessio and de Souza, Audrien Alves Andrade and Souza, Alessandra Beatriz Santos Rondon and de Souza, Ágatha Ferreira and Carvalho-Assef, Ana Paula D'Alincourt and Moreira, Aline Dos Santos and Dos Santos, Marcelo Trindade and Côrtes, Adriano Maurício de Almeida and Penna, Bruno de Araújo and Chagas, Thiago Pavoni Gomes and Aguiar-Alves, Fábio and da Silva, Fabrício Alves Barbosa},
  title = {Harnessing interpretable deep learning to predict resistance in Klebsiella pneumoniae},
  journal = {Frontiers in Cellular and Infection Microbiology},
  year = {2026},
  doi = {10.3389/fcimb.2026.1859508}
}
Kostenlos starten →