From genome to phenotype: advances, challenges and One Health perspectives in the genomic prediction of antimicrobial resistance
5 affiliations
- 1Laboratory for Process Engineering, Environment, Biotechnology and Energy (LEPABE), Associate Laboratory in Chemical Engineering (ALiCE), Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, Porto, Portugal
- 2National Institute for Agrarian and Veterinary Research (INIAV), Rua dos Lagidos, Vila do Conde, Portugal
- 3School of Medicine and Biomedical Sciences (ICBAS), University of Porto, Porto, Portugal
- 4Department of Chemistry and Biochemistry, Faculty of Sciences, Associated Laboratory for Green Chemistry (LAQV), Network of Chemistry and Technology (REQUIMTE), University of Porto, Porto, Portugal
- 5Centre of Biological Engineering (CEB), University of Minho, Braga, Portugal
Abstract
Whole-genome sequencing (WGS) has substantially improved the detection and surveillance of antimicrobial resistance (AMR) by enabling the rapid identification of acquired resistance genes, resistance-associated point mutations, and mobile genetic elements. However, the presence of a resistance determinant does not necessarily establish its functional contribution to the antimicrobial phenotype. This mini-review argues that a central limitation of current genotype-to-phenotype prediction is representational: multidimensional resistance architectures are frequently reduced to binary presence–absence profiles. Such representations disregard gene integrity, expression state, promoter architecture, copy number, genomic localization, regulatory topology, strain background, and within-isolate population heterogeneity. Using functionally silent determinants, pseudogenized resistance genes, inducible and dosage-dependent resistance, altered permeability and efflux, heteroresistance, and genomic-context dependence as examples, we examine how this loss of biological information can generate both false-resistant and false-susceptible predictions. We further discuss why metagenomic ARG detection cannot be directly equated with phenotypic resistance and why prediction models trained on restricted clinical populations may not generalize across species, lineages, and One Health compartments. Moving from determinant detection toward context-aware and quantitative phenotype prediction will require the selective integration of long-read sequencing, transcriptomics, copy-number analysis, population-aware sequencing, standardized phenotypic metadata, and externally validated computational models. The relevant unit of AMR prediction is therefore not the resistance gene alone, but its functional state within a specific genomic, cellular, population, and environmental context.
Related
- Integrated Analysis of Virulence, Antimicrobial Resistance, and Gastric Mucosal Phenotype in Helicobacter pylori from a High-Risk Mongolian Population
- Antimicrobial Therapy, Disinfection, Resistance Mechanisms, and Microbiota Interactions in Nontuberculous Mycobacteria: Current Evidence and Emerging Perspectives.
- Divergent stakeholder perspectives on antimicrobial resistance and One Health in Ghana: a qualitative study
- Concordance of next-generation sequencing and conventional microbiology for pathogen detection and antimicrobial resistance prediction in pneumonia
- Benchmarking nanopore-based strategies for antimicrobial resistance prediction
BibTeX
@article{gomes2026from,
title = {{From genome to phenotype: advances, challenges and One Health perspectives in the genomic prediction of antimicrobial resistance}},
author = {Emílio Gomes and Ricardo J. Figueiredo and João R. Mesquita and Carina Almeida and Daniela Araújo and Nuno F. Azevedo and Joana Castro},
journal = {Frontiers in Cellular and Infection Microbiology},
year = {2026},
month = oct,
doi = {10.3389/fcimb.2026.1924354},
url = {https://doi.org/10.3389/fcimb.2026.1924354},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - From genome to phenotype: advances, challenges and One Health perspectives in the genomic prediction of antimicrobial resistance
AU - Emílio Gomes
AU - Ricardo J. Figueiredo
AU - João R. Mesquita
AU - Carina Almeida
AU - Daniela Araújo
AU - Nuno F. Azevedo
AU - Joana Castro
PY - 2026
DA - 2026/10/21
JO - Frontiers in Cellular and Infection Microbiology
DO - 10.3389/fcimb.2026.1924354
UR - https://doi.org/10.3389/fcimb.2026.1924354
AB - Whole-genome sequencing (WGS) has substantially improved the detection and surveillance of antimicrobial resistance (AMR) by enabling the rapid identification of acquired resistance genes, resistance-associated point mutations, and mobile genetic elements. However, the presence of a resistance determinant does not necessarily establish its functional contribution to the antimicrobial phenotype. This mini-review argues that a central limitation of current genotype-to-phenotype prediction is representational: multidimensional resistance architectures are frequently reduced to binary presence–absence profiles. Such representations disregard gene integrity, expression state, promoter architecture, copy number, genomic localization, regulatory topology, strain background, and within-isolate population heterogeneity. Using functionally silent determinants, pseudogenized resistance genes, inducible and dosage-dependent resistance, altered permeability and efflux, heteroresistance, and genomic-context dependence as examples, we examine how this loss of biological information can generate both false-resistant and false-susceptible predictions. We further discuss why metagenomic ARG detection cannot be directly equated with phenotypic resistance and why prediction models trained on restricted clinical populations may not generalize across species, lineages, and One Health compartments. Moving from determinant detection toward context-aware and quantitative phenotype prediction will require the selective integration of long-read sequencing, transcriptomics, copy-number analysis, population-aware sequencing, standardized phenotypic metadata, and externally validated computational models. The relevant unit of AMR prediction is therefore not the resistance gene alone, but its functional state within a specific genomic, cellular, population, and environmental context.
SN - 2235-2988
ER -