Predicting cell-cell horizontal transmission of antibiotics resistance from genome and phenome

Lead Research Organisation: University of Birmingham
Department Name: Institute of Cancer and Genomic Sciences

Abstract

We propose to disclose candidate drug targets controlling the horizontal cell-cell transmission of anti-microbial resistance (AMR) and to predict AMR and its transmission dynamics from bacterial genome composition. We will integrate leading expertise from bacteriology, -omics and mathematical biology in the development of an integrated theoretical-empirical framework of plasmid borne transmission of AMR cassettes. We will employ massive-scale experimental evolution of Escherichia coli and Salmonella enterica gene deletion and overexpression collections, where adaptation requires transfer of AMR carrying conjugative plasmids. In addition, we will select for, identify and functionally dissect de novo mutations that promote horizontal transmission during long-term experimental evolution. Both approaches will disclose cellular functions controlling horizontal AMR transmission that are candidate targets for helper drugs delaying AMR development and spread. Second, we will sequence vast swaths of the genotype space inhabited by clinical bacterial isolates and disclose variants likely to alter transmission properties. DNA sequence data will be complemented by data on transcriptome, proteome and antibiotics resistance, allowing causally cohesive reconstruction of the history of antibiotics resistance. Third, we will integrate the omics data into a mathematical framework capable of predicting AMR transmission in clinical isolates, thereby laying the foundations for a future personalized medicine that tailors antibiotic choice to infection.

Technical Summary

We propose to disclose candidate drug targets controlling the horizontal cell-cell
transmission of anti-microbial resistance (AMR) and to predict AMR and its transmission
dynamics from bacterial genome composition. We will integrate leading expertise from
bacteriology, -omics and mathematical biology in the development of an integrated
theoretical-empirical framework of plasmid borne transmission of AMR cassettes. We will
employ massive-scale experimental evolution of Escherichia coli and Salmonella enterica
gene deletion and overexpression collections, where adaptation requires transfer of AMR
carrying conjugative plasmids. In addition, we will select for, identify and functionally dissect
de novo mutations that promote horizontal transmission during long-term experimental
evolution. Both approaches will disclose cellular functions controlling horizontal AMR
transmission that are candidate targets for helper drugs delaying AMR development and
spread. Second, we will sequence vast swaths of the genotype space inhabited by clinical
bacterial isolates and disclose variants likely to alter transmission properties. DNA sequence
data will be complemented by data on transcriptome, proteome and antibiotics resistance,
allowing causally cohesive reconstruction of the history of antibiotics resistance. Third, we
will integrate the omics data into a mathematical framework capable of predicting AMR
transmission in clinical isolates, thereby laying the foundations for a future personalized
medicine that tailors antibiotic choice to infection.

Planned Impact

The Center for AMR (CARe; http://care.gu.se/) at
University of Gothenburg (partners 1 and 2 are core members) provides an overall context
within which we will manage the project and communicate with industry, regulatory bodies
and patient organizations. CARe have direct channels to international and national
regulatory bodies, interacts with commercial actors on both the drug development and
diagnostics side and has well established interfaces with patient organizations such as
REACT (http://www.reactgroup.org/) on the international level and its Nordic (NordicAST;
http://www.nordicast.org/) and Swedish subsidiaries (STRAMA;
http://www.reactgroup.org/toolbox/the-swedish-strama-program-2/).
* To exploit the mathematical framework for predicting the spread of antibiotics
resistance (WP4), we will interact closely with 1928 diagnostics
(http://www.1928diagnostics.com/) and Nanoxis consulting
(http://nanoxisconsulting.com/). This is outlined in section 14.
* To exploit emerging candidate drug targets controlling horizontal transmission (WP1,
WP2), we will interact with medicinal and organic chemists within the Centre for
Antibiotics Resistance rEsearch (CARe). Partner 1 is deeply involved in setting-up a
research centre for the development of long-lasting antibiotics (DLLA) at the
University of Gothenburg. This centre will become central in proceeding from
candidate drug targets to candidate drugs. We also expect to hold explorative
discussions with EU-OPENSCREEN, EATRIS and ENABLE, as outlined in section
JPI-EC-AMR JTC 2016 Full-proposal application form

Publications

10 25 50
publication icon
Murphy R (2021) Genomic Epidemiology and Evolution of Escherichia coli in Wild Animals in Mexico. in mSphere

 
Title Automated scientific literature analysis on the epidemiology of antimicrobial resistance 
Description The manuscript concerns an automated text analysis pipeline to analyze trends of antimicrobials resistance level in the wealth of available textual data in the scientific literature. The manuscript will be submitted by June 2021. The manuscript will disclose a curated database of country-based trends of the rise of antimicrobials resistance over time, extracted from textual scientific data. A reproducible pipeline specifically designed for the analysis of AMR literature will be published as well. Details of the manuscript are as follows: Automated scientific literature analysis on the epidemiology of antimicrobial resistance Authors: Marie Lisandra Zepeda-Mendoza, Sam Benkwitz-Bedford*, Talip Yasir Demirtas, Willem van Schaik, Danesh Moradigaravand* *Funded by the grant Abstract Antimicrobial resistance (AMR) is a global health issue that is predicted to become a leading cause of death worldwide in the coming decades. While the scientific literature on AMR can be used to extract information on the main factors influencing the spread of AMR worldwide, manual literature reviews do not capture the wealth of information in the unstructured textual data. In this study, we identified the socio-economic factors that influence the presence of a given country in AMR-related scientific literature. To this end, we used machine learning and Natural Language Processing (NLP) methods to integrate scientific literature from PubMed with antibiotic consumption data and socio-economic data from different online sources. Our results reveal that both the consumption of antimicrobials and gross national income is the most explanatory features for the number of reports on AMR from each country, although the particular magnitude and order of the importance of each explanatory feature is different for each country. We also identified temporal trends in the use of antibiotic classes across different countries, which were concordant with the frequency of the country appearance in the literature. Moreover, we found that the rise of livestock antimicrobials consumption in a country was significantly linked with the rise of the AMR literature mentioning the country. This study is the first automated text analysis on AMR epidemiology and highlights the significance of public literature in inferring latent knowledge on the epidemiology of AMR which can be extracted by utilizing NLP techniques. The significance of the use of NLP and ML lays in the automation of text analysis to obtain insights that go beyond those that are obtained from manual meta-reviews. 
Type Of Material Data analysis technique 
Year Produced 2021 
Provided To Others? Yes  
Impact The paper presents the first application of Natural Language Processing in AMR research and a text classification pipeline to filter epidemiological studies. 
 
Title Genomic Epidemiology and Evolution of Escherichia coli in Wild Animals 
Description In this project, my group conducted an in-depth analysis of the genomic data set for a rare collection of E. coli from diverse sources. We reported genomic characteristics, pan-genome and database of virulence and AMR. 
Type Of Material Data analysis technique 
Year Produced 2021 
Provided To Others? Yes  
Impact Although wild animals are recognized as potential reservoirs for pathogenic E. coli strains, the knowledge of the population structure of E. coli in wild hosts is still scarce. In this study, we used the fine resolution of whole-genome sequencing to provide novel insights into the evolution of E. coli genomes from a small yet diverse collection of strains recovered within a broad range of wild animal species (including mammals and birds), the coevolution of E. coli strains with their hosts, and the genetics of pathogenicity of E. coli strains in wild hosts in Mexico. Our results provide evidence for the clinical importance of wild animals as reservoirs for pathogenic strains and highlight the need to include nonhuman hosts in the surveillance programs for E. coli infections. 
URL https://github.com/dmoradigaravand/WildHostEcoliMexico
 
Title Machine learning prediction of resistance to sub-inhibitory antimicrobial concentrations from Escherichia coli genomes 
Description The main results of the grant project will be published in the form of a major research paper, to be submitted by the end of March 2021. The research manuscript will be first to use machine learning algorithms to predict bacterial growth phenotypes from Whole Genome Sequencing (WGS) data under a wide range of concentrations of antimicrobials in Escherichia coli populations. The manuscript will disclose predictive biomarkers and a robust and reproducible predictive pipeline for predicting any quantitative bacterial trait from high throughput phenotypic assays. Details of the manuscript are as follows: Machine learning prediction of resistance to sub-inhibitory antimicrobial concentrations from Escherichia coli genomes Authors: Sam Benkwitz-Bedford*, Martin Palm, Talip Yasir Demirtas, Ville Mustonen, Anne Farewell, Jonas Warringer, Leopold Parts, Danesh Moradigaravand* *Funded by the grant Abstract: Escherichia coli is an important cause of bacterial infections worldwide, with multidrug-resistant strains incurring substantial costs on human lives. Besides therapeutic concentrations of antimicrobials in healthcare settings, the presence of sub-inhibitory antimicrobial residues in the environment and in the clinics selects for antimicrobial resistance (AMR), but the underlying biology is less well understood. We used machine learning to predict the population doubling time and yield of >1,400 genetically diverse E. coli expanding under exposure to three sub-inhibitory concentrations of six classes of antimicrobials from single nucleotide genetic variants, accessory gene variation and the presence of known AMR genes. We could predict cell yields in the held-out test data with an average correlation (Spearman's ?) of 0.63 (0.32 - 0.90 across concentrations) and cell doubling time with an average correlation of 0.47 (0.32 - 0.74 across concentrations), with moderate increases in sample size unlikely to improve predictions further. This points to the remaining missing heritability of growth under antibiotics exposure being explained by effects that are too rare or weak to be captured unless the sample size is dramatically increased, or by effects other than those conferred by the presence of individual SNPs and genes. Predictions based on whole-genome information were generally superior to those based only on known AMR genes, and also accurate for AMR resistance at therapeutic concentrations. We also presented genes and SNPs determining the predicted growth and thereby recapitulated the known AMR determinants. Finally, we estimated the effect sizes of resistance genes across the entire collection of strains, disclosing growth effects for known resistance genes for each strain. Our results underscore the potential of predictive modelling of growth patterns from genomic data under nontherapeutic sub-inhibitory concentrations of antimicrobials, although the remaining missing heritability poses an issue for achieving the accuracy and precision required for clinical use. 
Type Of Material Data analysis technique 
Year Produced 2021 
Provided To Others? Yes  
Impact The study is the first application of predictive models, based on machine learning, to predicting complex bacterial traits from genomic data in natural pathogen populations. 
 
Description Application of Natural Language Processing in Antimicrobial Resistance Research 
Organisation University of Birmingham
Department School of Immunity and Infection
Country United Kingdom 
Sector Academic/University 
PI Contribution I am leading a collaborative project to apply Natural Language Processing methods to understand the patterns of antimicrobial resistance epidemiology in the scientific literature. Collaborators include: Prof Dr. Willem van Schaik- School of Medical and Dental Sciences, Institute of Microbiology and Infection, University of Birmingham, Birmingham Dr Marie Lisandra Zepeda-Mendoza- School of Medical and Dental Sciences, Institute of Microbiology and Infection, University of Birmingham, Birmingham
Collaborator Contribution Dr Marie Lisandra Zepeda-Mendoza is involved in driving the analysis of datasets and methods conceptualization and development. Prof Dr. Willem van Schaik is involved in the analysis and interpretation of results.
Impact The collaboration resulted in a manuscript to be submitted by June 2021. Details of the manuscript are described in the research output section.
Start Year 2020
 
Description Machine Learning-aided characterisation of genotype-phenotype maps from large-scale chemical genomic assays 
Organisation University of Birmingham
Country United Kingdom 
Sector Academic/University 
PI Contribution An international collaboration between the chemical genomic lab, led by Dr Manuel Banzhaf, at the University of Birmingham. The collaboration aims at introducing machine learning and data mining methods in chemical genomics assays to identify the genetic factors underlying biofilm formation and improve functional annotation pipelines.
Collaborator Contribution Dr. Manuel Banzhaf leads the chemical genomics lab at the University of Birmingham. The lab conducts large-scale phenotypic assays in the form of forward and reverse genetic settings.
Impact Thus the following manuscript has been submitted, but we expect the collaboration to yield software, publications and datasets. The lipoprotein DolP affects cell separation in Escherichia coli, but not as an upstream regulator of NlpD Authors: Gabriela Boelter; Jack A. Bryant; Hannah Doherty; Peter Wotherspoon; Dema Alodaini; Xuyu Ma; Micheal Alao; Patrick Moynihan; Danesh Moradigaravand; Timothy J. Knowles; Monika Glinkowska; Ian R. Henderson; Manuel Banzhaf
Start Year 2021
 
Description The Joint Programming Initiative on Antimicrobial Resistance (JPI-AMR) 
Organisation The Wellcome Trust Sanger Institute
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution The MRC award was part of a wider European consortium in the Joint Programming Initiative on Antimicrobial Resistance (JPIAMR). The overall aim of the consortium was integrating Omics data from natural bacterial strains to predict major bacterial traits with clinical importance. The consortium was composed of PIs based in Sweden, Finland and the UK with the following specification: Prof. Dr. Anne Farewell- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Jonas Warringer- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Ville Mustonen- Professor- Organismal and Evolutionary Biology Research Programme, Department of Computer Science, Institute of Biotechnology, University of Helsinki, Helsinki, Finland. Leopold Parts- group Leader- Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridgeshire, United Kingdom. The contribution of my research team as part of the consortium was twofold: 1) To conduct a genomic epidemiology analysis on a diverse genomic collection of Escherichia coli strains retrieved from wild hosts to understand the clinical importance of non-clinical and environmental reservoirs (published manuscript 1). 2) To develop models to predict bacterial phenotypes from genomic data (manuscript 2 to be submitted in March 2021).
Collaborator Contribution Prof. Dr. Anne Farewell and Prof. Jonas Warringer produced the experimental data and provided the bacterial strains for the Whole Genome Sequencing (WGS). They were also actively involved in the manuscripts write-up and the analysis of the results for manuscripts 1 and 2. Prof. Ville Mustonen contributed to the analysis and interpretation of results for manuscripts 1 and 2. Dr. Leopald Parts coordinated the sequencing of strains at the Wellcome Sanger Institute. Furthermore, he actively contributed to the analysis and interpretation of results for manuscripts 1 and 2.
Impact Publication (1): Genomic Epidemiology and Evolution of Escherichia coli in Wild Animals in Mexico Robert Murphy, Martin Palm, Ville Mustonen, Jonas Warringer, Anne Farewell, Leopold Parts, Danesh Moradigaravand* 2021- mSphere- DOI: 10.1128/mSphere.00738-20 Manuscript (2) to be submitted in March 2021: Machine-learning prediction of resistance to sub-inhibitory antimicrobial concentrations from Escherichia coli genomes Authors: Sam Benkwitz-Bedford*, Martin Palm, Talip Yasir Demirtas, Ville Mustonen, Anne Farewell, Jonas Warringer, Leopold Parts, Danesh Moradigaravand* * Funded by the award Details of the manuscript are provided in the research output section.
Start Year 2019
 
Description The Joint Programming Initiative on Antimicrobial Resistance (JPI-AMR) 
Organisation University of Gothenburg
Department Department of Chemistry & Molecular Biology
Country Sweden 
Sector Academic/University 
PI Contribution The MRC award was part of a wider European consortium in the Joint Programming Initiative on Antimicrobial Resistance (JPIAMR). The overall aim of the consortium was integrating Omics data from natural bacterial strains to predict major bacterial traits with clinical importance. The consortium was composed of PIs based in Sweden, Finland and the UK with the following specification: Prof. Dr. Anne Farewell- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Jonas Warringer- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Ville Mustonen- Professor- Organismal and Evolutionary Biology Research Programme, Department of Computer Science, Institute of Biotechnology, University of Helsinki, Helsinki, Finland. Leopold Parts- group Leader- Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridgeshire, United Kingdom. The contribution of my research team as part of the consortium was twofold: 1) To conduct a genomic epidemiology analysis on a diverse genomic collection of Escherichia coli strains retrieved from wild hosts to understand the clinical importance of non-clinical and environmental reservoirs (published manuscript 1). 2) To develop models to predict bacterial phenotypes from genomic data (manuscript 2 to be submitted in March 2021).
Collaborator Contribution Prof. Dr. Anne Farewell and Prof. Jonas Warringer produced the experimental data and provided the bacterial strains for the Whole Genome Sequencing (WGS). They were also actively involved in the manuscripts write-up and the analysis of the results for manuscripts 1 and 2. Prof. Ville Mustonen contributed to the analysis and interpretation of results for manuscripts 1 and 2. Dr. Leopald Parts coordinated the sequencing of strains at the Wellcome Sanger Institute. Furthermore, he actively contributed to the analysis and interpretation of results for manuscripts 1 and 2.
Impact Publication (1): Genomic Epidemiology and Evolution of Escherichia coli in Wild Animals in Mexico Robert Murphy, Martin Palm, Ville Mustonen, Jonas Warringer, Anne Farewell, Leopold Parts, Danesh Moradigaravand* 2021- mSphere- DOI: 10.1128/mSphere.00738-20 Manuscript (2) to be submitted in March 2021: Machine-learning prediction of resistance to sub-inhibitory antimicrobial concentrations from Escherichia coli genomes Authors: Sam Benkwitz-Bedford*, Martin Palm, Talip Yasir Demirtas, Ville Mustonen, Anne Farewell, Jonas Warringer, Leopold Parts, Danesh Moradigaravand* * Funded by the award Details of the manuscript are provided in the research output section.
Start Year 2019
 
Description The Joint Programming Initiative on Antimicrobial Resistance (JPI-AMR) 
Organisation University of Helsinki
Country Finland 
Sector Academic/University 
PI Contribution The MRC award was part of a wider European consortium in the Joint Programming Initiative on Antimicrobial Resistance (JPIAMR). The overall aim of the consortium was integrating Omics data from natural bacterial strains to predict major bacterial traits with clinical importance. The consortium was composed of PIs based in Sweden, Finland and the UK with the following specification: Prof. Dr. Anne Farewell- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Jonas Warringer- Associate Professor- Department for Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden, Centre for Antibiotic Resistance Research at the University of Gothenburg, Gothenburg, Sweden. Prof. Dr. Ville Mustonen- Professor- Organismal and Evolutionary Biology Research Programme, Department of Computer Science, Institute of Biotechnology, University of Helsinki, Helsinki, Finland. Leopold Parts- group Leader- Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridgeshire, United Kingdom. The contribution of my research team as part of the consortium was twofold: 1) To conduct a genomic epidemiology analysis on a diverse genomic collection of Escherichia coli strains retrieved from wild hosts to understand the clinical importance of non-clinical and environmental reservoirs (published manuscript 1). 2) To develop models to predict bacterial phenotypes from genomic data (manuscript 2 to be submitted in March 2021).
Collaborator Contribution Prof. Dr. Anne Farewell and Prof. Jonas Warringer produced the experimental data and provided the bacterial strains for the Whole Genome Sequencing (WGS). They were also actively involved in the manuscripts write-up and the analysis of the results for manuscripts 1 and 2. Prof. Ville Mustonen contributed to the analysis and interpretation of results for manuscripts 1 and 2. Dr. Leopald Parts coordinated the sequencing of strains at the Wellcome Sanger Institute. Furthermore, he actively contributed to the analysis and interpretation of results for manuscripts 1 and 2.
Impact Publication (1): Genomic Epidemiology and Evolution of Escherichia coli in Wild Animals in Mexico Robert Murphy, Martin Palm, Ville Mustonen, Jonas Warringer, Anne Farewell, Leopold Parts, Danesh Moradigaravand* 2021- mSphere- DOI: 10.1128/mSphere.00738-20 Manuscript (2) to be submitted in March 2021: Machine-learning prediction of resistance to sub-inhibitory antimicrobial concentrations from Escherichia coli genomes Authors: Sam Benkwitz-Bedford*, Martin Palm, Talip Yasir Demirtas, Ville Mustonen, Anne Farewell, Jonas Warringer, Leopold Parts, Danesh Moradigaravand* * Funded by the award Details of the manuscript are provided in the research output section.
Start Year 2019
 
Description United Arab Emirates AMR consortium 
Organisation Khalifa University
Country United Arab Emirates 
Sector Academic/University 
PI Contribution Currently, I am the program director at the University of Birmingham, Dubai campus for their MSc course in Bioinformatics. As part of this role, I have extensively engaged and developed a regional network for AMR research. A national-wide consortium, composed of five major universities in the UAE,on AMR was set up in the UAE as a result of the efforts. AMR epidemiology is not well studied in the broader Middle East, so the consortium's aim is to introduce precision epidemiology and machine learning models to dissect and predict AMR in the UAE and broader Middle East.
Collaborator Contribution As the lead researcher, I will analyze the genetic data, oversee and disseminate the findings through publications and presentations. Prof. Dean Everett, from Khalifa University, and Prof. Abiola, from MBR University, will lead the initiative, manage sample collection from UAE hospitals and metadata curation and sequencing.
Impact The collaboration will result in multiple papers and reports on AMR epidemiology in the Middle East.
Start Year 2022
 
Description United Arab Emirates AMR consortium 
Organisation Mohammed Bin Rashid University Of Medicine and Health Sciences
Country United Arab Emirates 
Sector Academic/University 
PI Contribution Currently, I am the program director at the University of Birmingham, Dubai campus for their MSc course in Bioinformatics. As part of this role, I have extensively engaged and developed a regional network for AMR research. A national-wide consortium, composed of five major universities in the UAE,on AMR was set up in the UAE as a result of the efforts. AMR epidemiology is not well studied in the broader Middle East, so the consortium's aim is to introduce precision epidemiology and machine learning models to dissect and predict AMR in the UAE and broader Middle East.
Collaborator Contribution As the lead researcher, I will analyze the genetic data, oversee and disseminate the findings through publications and presentations. Prof. Dean Everett, from Khalifa University, and Prof. Abiola, from MBR University, will lead the initiative, manage sample collection from UAE hospitals and metadata curation and sequencing.
Impact The collaboration will result in multiple papers and reports on AMR epidemiology in the Middle East.
Start Year 2022