Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility

Anand V. Sastry; Nicholas Dillon; Amitesh Anand; Saugat Poudel; Ying Hefner; Sibei Xu; Richard Szubin; Adam M. Feist; Victor Nizet; Bernhard Palssona

doi:10.1128/mSphere.00443-21

Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility

Anand V. Sastry, Nicholas Dillon, Amitesh Anand, Saugat Poudel, Ying Hefner, Sibei Xu, Richard Szubin, Adam M. Feist, Victor Nizet, Bernhard Palssona^*

^*Corresponding author for this work

University of California at San Diego

Research output: Contribution to journal › Journal article › Research › peer-review

145 Downloads (Pure)

Abstract

In vitro antibiotic susceptibility testing often fails to accurately predict in vivo drug efficacies, in part due to differences in the molecular composition between standardized bacteriologic media and physiological environments within the body. Here, we investigate the interrelationship between antibiotic susceptibility and medium composition in Escherichia coli K-12 MG1655 as contextualized through machine learning of transcriptomics data. Application of independent component analysis, a signal separation algorithm, shows that complex phenotypic changes induced by environmental conditions or antibiotic treatment are directly traced to the action of a few key transcriptional regulators, including RpoS, Fur, and Fnr. Integrating machine learning results with biochemical knowledge of transcription factor activation reveals medium-dependent shifts in respiration and iron availability that drive differential antibiotic susceptibility. By extension, the data generation and data analytics workflow used here can interrogate the regulatory state of a pathogen under any measured condition and can be applied to any strain or organism for which sufficient transcriptomics data are available.

Original language	English
Journal	mSphere
Volume	6
Issue number	4
Pages (from-to)	1-16
Number of pages	16
ISSN	1535-9786
DOIs	https://doi.org/10.1128/mSphere.00443-21
Publication status	Published - Aug 2021

Bibliographical note

Funding Information:
This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under contract DE-AC02-05CH11231. This study was funded by the Novo Nordisk Foundation Center for Biosustainability and the Technical University of Denmark (NNF10CC1016517) and by the NIH NIAID (1-U01-AI124316). Nick Dillon was additionally supported by grant T32-NIH-5T32HD087978-05. We declare there are no competing interests.

Publisher Copyright:
© 2021. Sastry et al. This is an openaccess article distributed under the terms of the Creative Commons Attribution 4.0 International license.

Keywords

Antibiotics
Independent component analysis
Iron regulation
Machine learning
RNA-seq
Transcriptional regulation

Access to Document

10.1128/mSphere.00443-21

FulltextFinal published version, 1.96 MB

OpenUrl availability

Full text

Cite this

@article{5fce55cddd60402597705f1b36db3ae5,

title = "Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility",

abstract = "In vitro antibiotic susceptibility testing often fails to accurately predict in vivo drug efficacies, in part due to differences in the molecular composition between standardized bacteriologic media and physiological environments within the body. Here, we investigate the interrelationship between antibiotic susceptibility and medium composition in Escherichia coli K-12 MG1655 as contextualized through machine learning of transcriptomics data. Application of independent component analysis, a signal separation algorithm, shows that complex phenotypic changes induced by environmental conditions or antibiotic treatment are directly traced to the action of a few key transcriptional regulators, including RpoS, Fur, and Fnr. Integrating machine learning results with biochemical knowledge of transcription factor activation reveals medium-dependent shifts in respiration and iron availability that drive differential antibiotic susceptibility. By extension, the data generation and data analytics workflow used here can interrogate the regulatory state of a pathogen under any measured condition and can be applied to any strain or organism for which sufficient transcriptomics data are available.",

keywords = "Antibiotics, Independent component analysis, Iron regulation, Machine learning, RNA-seq, Transcriptional regulation",

author = "Sastry, {Anand V.} and Nicholas Dillon and Amitesh Anand and Saugat Poudel and Ying Hefner and Sibei Xu and Richard Szubin and Feist, {Adam M.} and Victor Nizet and Bernhard Palssona",

note = "Funding Information: This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under contract DE-AC02-05CH11231. This study was funded by the Novo Nordisk Foundation Center for Biosustainability and the Technical University of Denmark (NNF10CC1016517) and by the NIH NIAID (1-U01-AI124316). Nick Dillon was additionally supported by grant T32-NIH-5T32HD087978-05. We declare there are no competing interests. Publisher Copyright: {\textcopyright} 2021. Sastry et al. This is an openaccess article distributed under the terms of the Creative Commons Attribution 4.0 International license.",

year = "2021",

month = aug,

doi = "10.1128/mSphere.00443-21",

language = "English",

volume = "6",

pages = "1--16",

journal = "mSphere",

issn = "1535-9786",

publisher = "American Society for Microbiology",

number = "4",

}

TY - JOUR

T1 - Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility

AU - Sastry, Anand V.

AU - Dillon, Nicholas

AU - Anand, Amitesh

AU - Poudel, Saugat

AU - Hefner, Ying

AU - Xu, Sibei

AU - Szubin, Richard

AU - Feist, Adam M.

AU - Nizet, Victor

AU - Palssona, Bernhard

N1 - Funding Information: This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under contract DE-AC02-05CH11231. This study was funded by the Novo Nordisk Foundation Center for Biosustainability and the Technical University of Denmark (NNF10CC1016517) and by the NIH NIAID (1-U01-AI124316). Nick Dillon was additionally supported by grant T32-NIH-5T32HD087978-05. We declare there are no competing interests. Publisher Copyright: © 2021. Sastry et al. This is an openaccess article distributed under the terms of the Creative Commons Attribution 4.0 International license.

PY - 2021/8

Y1 - 2021/8

N2 - In vitro antibiotic susceptibility testing often fails to accurately predict in vivo drug efficacies, in part due to differences in the molecular composition between standardized bacteriologic media and physiological environments within the body. Here, we investigate the interrelationship between antibiotic susceptibility and medium composition in Escherichia coli K-12 MG1655 as contextualized through machine learning of transcriptomics data. Application of independent component analysis, a signal separation algorithm, shows that complex phenotypic changes induced by environmental conditions or antibiotic treatment are directly traced to the action of a few key transcriptional regulators, including RpoS, Fur, and Fnr. Integrating machine learning results with biochemical knowledge of transcription factor activation reveals medium-dependent shifts in respiration and iron availability that drive differential antibiotic susceptibility. By extension, the data generation and data analytics workflow used here can interrogate the regulatory state of a pathogen under any measured condition and can be applied to any strain or organism for which sufficient transcriptomics data are available.

AB - In vitro antibiotic susceptibility testing often fails to accurately predict in vivo drug efficacies, in part due to differences in the molecular composition between standardized bacteriologic media and physiological environments within the body. Here, we investigate the interrelationship between antibiotic susceptibility and medium composition in Escherichia coli K-12 MG1655 as contextualized through machine learning of transcriptomics data. Application of independent component analysis, a signal separation algorithm, shows that complex phenotypic changes induced by environmental conditions or antibiotic treatment are directly traced to the action of a few key transcriptional regulators, including RpoS, Fur, and Fnr. Integrating machine learning results with biochemical knowledge of transcription factor activation reveals medium-dependent shifts in respiration and iron availability that drive differential antibiotic susceptibility. By extension, the data generation and data analytics workflow used here can interrogate the regulatory state of a pathogen under any measured condition and can be applied to any strain or organism for which sufficient transcriptomics data are available.

KW - Antibiotics

KW - Independent component analysis

KW - Iron regulation

KW - Machine learning

KW - RNA-seq

KW - Transcriptional regulation

U2 - 10.1128/mSphere.00443-21

DO - 10.1128/mSphere.00443-21

M3 - Journal article

C2 - 34431696

AN - SCOPUS:85114968931

SN - 1535-9786

VL - 6

SP - 1

EP - 16

JO - mSphere

JF - mSphere

IS - 4

ER -

Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility

Abstract

Bibliographical note

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this