Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


41641.pdf (2.95 MB)

ETD Abstract Container

Abstract Header

Evaluating the Bacterial (meta)genome for Antimicrobial Resistance using High-throughput Sequencing

Van Camp, Pieter-Jan
http://orcid.org/0000-0003-0675-2011
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1649771158512402

Abstract Details

2022, PhD, University of Cincinnati, Medicine: Biomedical Informatics.
Bacterial antimicrobial resistance (AMR) is one of the major challenges 21st century medicine is facing with multidrug-resistant strains causing millions of deaths worldwide every year. Traditional clinical laboratory evaluation by culturing bacterial pathogens and subsequent AMR testing is effective but relatively slow and expensive compared to the emerging novel approach of using next-generation sequencing (NGS). Furthermore, standard clinical pipelines (i.e., used in hospital setting) are also limited to evaluation of a single bacterial species and fail to comprehensively analyze more complex samples as might originate from a microbiome environment. This work outlines novel biomedical informatics approaches to evaluating and understanding AMR using next-generation sequencing data of both isolate and metagenomic origin. First, we developed a new machine learning approach to leveraging NGS for predicting complex genotype to phenotype relationships as found in a set of clinically important Gram-negative bacteria. The XGBoost, decision-tree based, AMR prediction models demonstrate high predictive capabilities. By extracting the most important features used in the decision-making process from the models we further validated the predictions by showing biological relevance of the used antimicrobial resistance genes and their connections to specific antibiotics or relationship with other relevant genes. Using R Shiny, a demo application was developed to provide an interface to the prediction results underscoring the importance of tailoring output to the target audience. The work is then extended into the realm of metagenomics. A novel tool SEQ2MGS was built to quickly generate realistic benchmarking data. By mixing reads from existing NGS experiments, (semi)annotated datasets can be created and used for validating metagenomics pipelines. We have subsequently developed such pipeline dubbed MGS2AMR and described in the last part of this work. It takes metagenomics sequencing experiments and performs a set of steps, including targeted gene assembly, assembly graph evaluation, alignment to known bacterial genomes, and grouping of bacteria and antimicrobial resistance genes into clusters. MGS2AMR effectively identifies the pathogenic bacteria from the background metagenome to allow subsequent detailed AMR analysis, such as genotype to phenotype prediction as presented in the first part of this work. The research and new bioinformatics tools presented in this study demonstrate the immense potential of using NGS data for the fast and accurate evaluations of AMR in the bacterial (meta)genome. Our work sets the stage for better antibiotics management in the clinical setting whilst improving patient outcomes and encouraging more targeted use of antibiotics to slow down the emergence of resistance to antibiotics.
Alexey Porollo, Ph.D. (Committee Member)
Danny T. Y. Wu, PhD (Committee Member)
Surya Prasath, Ph.D. (Committee Member)
David Haslam, M.D. (Committee Member)
162 p.
Bioinformatics
metagenomics; biomedical informatics; antimicrobial resistance; genome sequencing; machine learning

Recommended Citations

Citations

  • Van Camp, P.-J. (2022). Evaluating the Bacterial (meta)genome for Antimicrobial Resistance using High-throughput Sequencing [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1649771158512402

    APA Style (7th edition)

  • Van Camp, Pieter-Jan. Evaluating the Bacterial (meta)genome for Antimicrobial Resistance using High-throughput Sequencing. 2022. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1649771158512402.

    MLA Style (8th edition)

  • Van Camp, Pieter-Jan. "Evaluating the Bacterial (meta)genome for Antimicrobial Resistance using High-throughput Sequencing." Doctoral dissertation, University of Cincinnati, 2022. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1649771158512402

    Chicago Manual of Style (17th edition)

ucin1649771158512402
90
© 2022, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.