Tackling a hospital superbug with genomics and metabolic modelling

Klebsiella pneumoniae is a major cause of antibiotic resistant hospital-associated infections that can be extremely difficult to treat. The World Health Organization has identified this bacterium as a high priority pathogen for which novel control strategies are urgently required. But there are key gaps in our knowledge about the underlying biology of this organism that limit our ability to design and implement these control strategies. Over the past three years, my lab has been exploring a new approach for understanding Klebsiella biology and population diversity, by combining comparative genomics with genome-scale metabolic modelling. We have generated metabolic models for thousands of K. pneumoniae and predicted their ability to grow in more than 1000 distinct growth conditions. This work has shown that there is a huge amount of metabolic diversity among K. pneumoniae strains, including traits that are known to play a role in colonisation of the human gut (from which the vast majority of infections in hospital patients are thought to emerge) and those known to directly influence the progression of disease. However, a major caveat of our current analyses is that the metabolic models do not consider some of the key metabolic processes that are known to be important during infection, such as iron metabolism, and the production of cell surface sugars that protect the bacteria from the human immune system. In this project the student will contribute to updating our reference metabolic model to address the caveats described above. He/she/they will then use the updated reference to generate individual models for large collections of K. pneumoniae and simulate growth in a range of conditions, in order to further explore the population diversity, identify metabolic virulence determinants and/or metabolic choke-points that serve as novel drug targets. This work is primarily computational, spanning genomics and metabolic modelling, but could optionally include a small amount of laboratory work to validate model predictions. The scope of the work can be refined to accommodate projects of varying length and is best suited for students interested in the application of computational biology approaches (including command-line programs) to analyse and interpret large datasets. Prior experience using the Unix operating system and the Python programming language is preferred but not essential.