Advancing detection of complex antimicrobial resistance mechanisms in bacterial pathogens

The burgeoning pandemic of antimicrobial resistant (AMR) bacterial pathogens poses a major threat to global health. In Australia alone, 6.5 deaths per 100,000 population are caused by AMR bacteria and these infections often occur in vulnerable patients in hospital settings. Much of this AMR burden is hidden to clinicians as current phenotypic methods for diagnosing resistance are slow and do not reveal its mechanism or mode of acquisition: pieces of information that are vital for managing and preventing AMR infections and outbreaks in the long term. Whole-genome sequencing (WGS) is becoming routine for surveillance of AMR bacterial pathogens and there is an urgent need to predict AMR phenotypes using genomic data. Extended-spectrum beta-lactamase (ESBL) and carbapenemase producing Enterobacteriaceae are priority-1 organisms. Resistance to beta-lactam antibiotics is mostly due to the production of beta-lactamases, and drugs comprising a combination of beta-lactamase inhibitor and a beta-lactam antibiotic are used to negate the effect of some beta-lactamases. Prediction of resistance to beta-lactams using WGS data is more difficult than most other drug classes, as the specific gene allele and expression levels can influence the observed phenotype – e.g., bacteria that over-express beta-lactamases are able to escape treatment with more sophisticated beta-lactams and beta-lactam/inhibitor combinations. Gene duplication is one way a bacterial isolate can over-express beta-lactamases. Current tools used to detect AMR-encoding genes typically only identify the presence or absence of a specific gene, but not whether the gene is duplicated. In this project, the student will investigate clinical genome data to identify genomes with duplicated beta-lactamase genes. They will compare this with the phenotypic results for relevant beta-lactam antimicrobials, to determine the associations between specific beta-lactamase genes and their resulting phenotype. New genomic methods will be developed to better characterise gene duplication from WGS data. This project is best suited for students interested in the application of computational biology approaches (including command-line programs) to analyse and interpret large datasets. Specific analysis approaches will include de novo genome assembly and annotation, reference-based variant detection, BLAST search, and methods development. If desired there is also scope for the student to undertake additional validation of their findings in a wet-lab laboratory, utilising qPCR and broth microdilution.