Predicting Chemical Stressor Effects on Conjugative Plasmid Transfer Rates Using Machine Learning.
- 1Faculty of Science and Technology, Department of Biochemistry and Molecular Biology, Euskal Herriko Unibertsitatea (EHU), 48940Leioa, Bizkaia, Spain.
- 2Faculty of Hydrosciences, Institute of Hydrobiology, Technische Universität Dresden, 01217Dresden, Saxony, Germany.
Abstract
Chemical stressors in environmental and clinical settings can alter conjugative plasmid transfer, which is a key process driving antibiotic resistance dissemination. However, predicting these effects across chemically diverse compounds is challenging. Here, we compiled and curated data on the effect of chemical stressors on RP4 plasmid transfer, extracting compound identity, concentration, conjugation frequency, and experimental metadata from 73 in vitro studies. Chemicals were characterized by functional use categories and molecular descriptors, including physicochemical properties, MACCS keys, Morgan fingerprints, and extended descriptors from the OCHEM platform (AlvaDesc, Dragon 7, and 3D-shape descriptors). An exhaustive evaluation of 1022 feature set combinations across 3066 models showed that chemical concentration alone provides no meaningful predictive signal (Q2ext <0), whereas integrating Morgan fingerprints substantially improves predictive performance. The best-performing model combined concentration, Morgan fingerprints, and experimental conditions using extreme gradient boosting (XGB), achieving Q2ext = 0.373 on a held-out compound-aware test set and a median Q2ext of 0.319 within the defined applicability domain across 20 independent splits. SHAP analysis revealed that Morgan fingerprints account for 69% of total predictive weight, with quaternary ammonium compounds and long-chain alkyl structures associated with inhibition of conjugation and sulfonamide and amine-containing compounds associated with promotion. These results demonstrate that chemical structural features provide critical information for predicting the chemical modulation of conjugative plasmid transfer. This work highlights that combining chemically informed descriptors with machine learning (ML) anticipates chemical contributions to antibiotic resistance dissemination, providing a scalable framework for the prioritization and risk assessment of chemicals.
Related
- Predicting the Elastic Properties of a Cemented Granular Material during Chemical Damage (Debonding)
- Morphological, molecular, and chemical characters support the transfer of the tropical demosponge genus Rhabdastrella to the family Geodiidae
- A source-target transfer learning framework with physics priors for subsurface chlorophyll reconstruction in the Northwest Pacific Ocean
- Defining and predicting groundwater drought dynamics in a temperate maritime region using Standardised Groundwater Index (SGI) and ensemble machine learning
- Predicting freeway incident duration using machine learning
BibTeX
@article{reysogo2026predicting,
title = {{Predicting Chemical Stressor Effects on Conjugative Plasmid Transfer Rates Using Machine Learning.}},
author = {Ana Rey-Sogo and David Kneis and Alan Xavier Elena and Uli Klümper and Itziar Alkorta},
journal = {Environmental science \& technology},
year = {2026},
month = oct,
doi = {10.1021/acs.est.6c02828},
url = {https://europepmc.org/article/MED/42916015},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Predicting Chemical Stressor Effects on Conjugative Plasmid Transfer Rates Using Machine Learning.
AU - Ana Rey-Sogo
AU - David Kneis
AU - Alan Xavier Elena
AU - Uli Klümper
AU - Itziar Alkorta
PY - 2026
DA - 2026/10/01
JO - Environmental science & technology
DO - 10.1021/acs.est.6c02828
UR - https://europepmc.org/article/MED/42916015
AB - Chemical stressors in environmental and clinical settings can alter conjugative plasmid transfer, which is a key process driving antibiotic resistance dissemination. However, predicting these effects across chemically diverse compounds is challenging. Here, we compiled and curated data on the effect of chemical stressors on RP4 plasmid transfer, extracting compound identity, concentration, conjugation frequency, and experimental metadata from 73 in vitro studies. Chemicals were characterized by functional use categories and molecular descriptors, including physicochemical properties, MACCS keys, Morgan fingerprints, and extended descriptors from the OCHEM platform (AlvaDesc, Dragon 7, and 3D-shape descriptors). An exhaustive evaluation of 1022 feature set combinations across 3066 models showed that chemical concentration alone provides no meaningful predictive signal (Q2ext <0), whereas integrating Morgan fingerprints substantially improves predictive performance. The best-performing model combined concentration, Morgan fingerprints, and experimental conditions using extreme gradient boosting (XGB), achieving Q2ext = 0.373 on a held-out compound-aware test set and a median Q2ext of 0.319 within the defined applicability domain across 20 independent splits. SHAP analysis revealed that Morgan fingerprints account for 69% of total predictive weight, with quaternary ammonium compounds and long-chain alkyl structures associated with inhibition of conjugation and sulfonamide and amine-containing compounds associated with promotion. These results demonstrate that chemical structural features provide critical information for predicting the chemical modulation of conjugative plasmid transfer. This work highlights that combining chemically informed descriptors with machine learning (ML) anticipates chemical contributions to antibiotic resistance dissemination, providing a scalable framework for the prioritization and risk assessment of chemicals.
SN - 0013-936X
SN - 1520-5851
ER -