Truace
Subscribe

Advances in machine learning and deep learning for antimicrobial resistance prediction

Antimicrobial resistance (AMR) is a major public health risk. Testing antimicrobial susceptibility in patients using conventional methods takes a week or more to deliver results, advances in artificial intelligence (AI), including machine learning (ML), and deep learning (DL), have created an opportunity to predict AMR with speed by combining data from different sources such as genomic, transcriptomic, proteomic, phenotypic, and clinical information. Recent advances in whole genome sequencing, metagenomic method…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-2102· cite & verify
Advances in machine learning and deep learning for antimicrobial resistance prediction
"Zonal ventilation model for the airborne spread of bacteria within a microbiology laboratory" by Jonathan Carruthers, Martín López-García, Joseph J. Gillard, Thomas R. Laws, Grant Lythe,Carmen Molina-París is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/.

In brief

This peer-reviewed review surveys how machine learning and deep learning are used to predict antimicrobial resistance, a major public health risk. It covers the data types, feature engineering approaches and algorithms used in studies of single pathogens and across many species.

The work matters because conventional susceptibility testing is slow, and faster prediction could support treatment decisions, monitoring and drug development. The authors frame the benefits as potential rather than proven at scale, and note that further validation is still needed before these models can be relied on.

Main points

  1. Conventional antimicrobial susceptibility testing in patients takes a week or more to return results.
  2. Models draw on genomic, transcriptomic, proteomic, phenotypic and clinical data.
  3. Whole genome sequencing, metagenomics, MALDI-TOF mass spectroscopy and the CARD database are described as essential components.
  4. Studies have used ML and DL to identify resistance genes, find new AMR determinants and predict AMR trends.
  5. Listed challenges include dataset diversity, class imbalance, inconsistent testing methods and limited interpretability.

The gain

Machine learning and deep learning can combine genomic, proteomic, phenotypic and clinical data to predict antimicrobial resistance faster than conventional susceptibility testing that takes a week or more.

The rundown

The review explains that advances in sequencing and mass spectroscopy, together with curated resources such as the Comprehensive Antibiotic Resistance Database, provide the foundation for robust prediction models. It reports that many studies have used ML and DL to detect resistance genes and discover new resistance determinants.

Beyond detection, the cited studies are described as using these methods to forecast resistance trends, improve antimicrobial stewardship and aid drug discovery. The review assesses methods in both pathogen-specific and pan-species work.

Future directions include explainable AI models, federated learning to improve datasets, and integration of digital diagnostics into electronic health records. The authors conclude these tools could advance precision medicine, AMR monitoring and drug development.

Sources

  1. Peer-reviewedDiscover Applied Sciences2026-10-20

The debate