Evaluation of a machine learning system for genomic antimicrobial susceptibility determination on a clinically representative test set
Microbiology Spectrum · 2026
DOI: 10.1128/spectrum.00564-25 ↗Kurz gefasst
Evaluationsstudie in der Zeitschrift Microbiology Spectrum, 2026. Verfasst von Jason D Wittenbach und 17 weiteren Autor:innen. Beteiligte Einrichtung: Day Zero Diagnostics, Inc, Watertown, Massachusetts, USA. Hauptthemen laut MEDLINE: Machine Learning, Anti-Bacterial Agents, Bacteria, Genomics. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator.
Abstract (Originalsprache)
Next-generation sequencing is anticipated to transform infectious disease treatment by enabling faster diagnosis, yet a clinical-grade method for genomic AST determination is still lacking. We assessed the accuracy of Keynome g AST, a k -mer based machine learning system for genomic AST, on 956 clinical bacterial isolates. We compared 7,801 predictions of susceptible (S), intermediate (I), and resistant (R) across 97 species-drug combinations against phenotypic AST results. Across the full data set, the Keynome g AST Qualified panel achieved a categorical agreement of 96.9% with very major and major error rates of 1.4% and 0.8%, respectively. At the level of individual species-drug combinations with sufficient data to assess, performance was comparable to the aggregate, with median categorical agreement of 97.4%. We also compared the performance of Keynome g AST to ResFinder, a simple resistance marker approach, in distinguishing susceptible vs non-susceptible samples. On the full data set, Keynome g AST performance was significantly better than that of ResFinder (binary accuracy of 96.0% vs 83.5%). When analyzed at the species-drug level, ResFinder’s poor performance was seen to be driven by either low sensitivity due to phenotypically resistant samples lacking resistance markers (18.0% of such samples) or low specificity due to the presence of resistance markers in phenotypically susceptible samples (15.1% of such samples). In total, these results demonstrate the dual advantages of machine learning and whole-genome profiling that Keynome g AST has over simple resistance marker approaches, enabling high accuracy genomic AST across a range of clinically relevant species-drug combinations. IMPORTANCE Microbial genome sequencing presents an exciting opportunity for rapid diagnosis of infectious diseases, but interpreting the resulting data for clinical use remains a challenge. We report a new machine learning method that predicts a bacterial strain's antibiotic resistance profile based solely on its genomic sequence. This method could lead to new, faster diagnostic tools that quickly identify the most effective antibiotic therapy.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Day Zero Diagnostics, Inc, Watertown, Massachusetts, USA.
- Department of Pathology and Laboratory Medicine, University of California Irvine Health, Orange, California, USA.
- Epidemiology and Infection Prevention, University of California Irvine Health, Irvine, California, USA.
ORCID-Profile: Jason D Wittenbach, Arolyn Conwill, Miriam H Huntley
Förderung
- Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator
Zitieren
APA
Wittenbach, J. D., Conwill, A., Sansum, H., Gassett, A., Gardner, A., Brookhart, A., Hollowell, T., Knysh, P., Worley, N. B., Billings, N., Herriott, I. C., Shimabukuro, J. A., Quan, K. A., Madey, K. M., Huang, S. S., Sater, M. R. A., Bittencourt, C. E., & Huntley, M. H. (2026). Evaluation of a machine learning system for genomic antimicrobial susceptibility determination on a clinically representative test set. Microbiology Spectrum. https://doi.org/10.1128/spectrum.00564-25
Vancouver
Wittenbach JD, Conwill A, Sansum H, Gassett A, Gardner A, Brookhart A, et al. Evaluation of a machine learning system for genomic antimicrobial susceptibility determination on a clinically representative test set. Microbiology Spectrum. 2026. doi:10.1128/spectrum.00564-25
BibTeX
@article{Wittenbach2026,
author = {Wittenbach, Jason D and Conwill, Arolyn and Sansum, Hayden and Gassett, Alison and Gardner, Adam and Brookhart, Allison and Hollowell, Talia and Knysh, Paul and Worley, Nicholas B and Billings, Nicole and Herriott, Ian C and Shimabukuro, Julie A and Quan, Kathleen A and Madey, Keith M and Huang, Susan S and Sater, Mohamad R A and Bittencourt, Cassiana E and Huntley, Miriam H},
title = {Evaluation of a machine learning system for genomic antimicrobial susceptibility determination on a clinically representative test set},
journal = {Microbiology Spectrum},
year = {2026},
doi = {10.1128/spectrum.00564-25}
}
EUVIMED