
Antimicrobial resistance (AMR) occurs when bacteria and other microbes that cause infections like pneumonia or tuberculosis gain the ability to resist treatment by antibiotics or other antimicrobial medicines.
Bacteria evolve rapidly, which means there can be thousands of new generations of bacteria with antimicrobial resistance genes within a week. Bacteria can become resistant to antibiotics in multiple ways, including random mutations in their genetic code and over-exposure to antibiotics.
Dr Elizabeth ‘Lilly’ Cummins is a Postdoctoral Research Associate at the Ineos Oxford Institute for antimicrobial research (IOI) Sheppard Lab, examining how evolutionary changes in genomes allow bacteria to become resistant to antibiotics.
Her work focuses on Escherichia coli (E.coli), a World Health Organisation designated ESKAPE pathogen, meaning infections have reduced treatment options and increased death rates due to treatment failure. This year the UK Health Security Agency reported an E.coli outbreak caused by contaminated pre-packaged sandwiches that led to the death of two people and the hospitalisation of 122.
Whilst much work has been done to establish which bacterial genes make a pathogen resistant to a specific antibiotic, it is unclear how some pathogenic lineages of E.coli acquire resistance whilst others do not. There are certain bacterial genetic backgrounds that can increase the likelihood of evolving resistance. Lilly focuses on the evolutionary events that lead to a resistance gene being acquired.
My work begins at the start of the evolutionary timeline, focusing on the events prior to resistance occurring in pathogens. I want to know why certain lineages within the E.coli species have a higher chance of evolving resistance to antibiotics. This will inform clinical practices to tailor antibiotic treatments, for example if a specific lineage is likely to gain resistance to a certain antibiotic, an alternative treatment plan could be recommended.

Lilly’s focus is on genetic import in E.coli that is associated with multidrug-resistant (MDR) sequence types. Sequence types (STs) are assigned based on Multi-Locus Sequence Typing (MLST), categorising bacteria based on 7 genes within the species.
Lilly’s project explores why some sequence types of bacteria develop resistance. Data show that some sequence types have persisted in clinical settings for a long time without evolving high levels of resistance, whereas other sequence types that have been recently described in clinical settings have much higher levels of resistance, being resistant to multiple antibiotics.
Bacteria can evolve resistance in multiple ways, including single mutations that result in resistance to an antibiotic. A major driver of resistance is through the acquisition of plasmids, small circular pieces of DNA that are exchanged between neighbouring bacteria. Some plasmids confer resistance to lots of different antibiotics, called multidrug-resistance (MDR) plasmids, whilst others only confer resistance to one antibiotic.
Research shows that bacteria which already have acquired a certain plasmid may be unable to acquire other MDR plasmids, so acquisition of resistance to several antibiotics can be blocked, maintaining the effectiveness of a wider range of treatments for that pathogen. There is potential to exploit the naturally occurring plasmid pathway to prevent bacteria acquiring resistance in the first place.
As part of Professor Sam Sheppard’s group, Lilly works in a ‘dry lab’, applying computational analyses to understand real-world populations of bacteria from a range of settings. This is done using data from naturally occurring bacteria as opposed to genetically engineered lab strains of bacteria. Data can vary in geography and timeframe, with the earliest data from the 1980s, when genomic data was first recorded on a large scale.
Much of this data comes from PubMLST, an open-access resource of microbial genomic data, which is maintained by the lab teams of Prof Martin Maiden and Prof Sam Sheppard, both based in the Department of Biology, University of Oxford.
Lilly’s team aim to be able to use bacterial genome sequencing to forecast resistance. “Our goal is to sequence a pathogen’s genome and then assess whether it has the potential to become resistant. This is useful because it means that the pathogen could be channelled into evolutionary dead-ends, stopping them from developing resistance – so preventing resistance from occurring in the first place rather than finding cures for drug-resistant infections.”
Our goal is to sequence a pathogen’s genome and then assess whether it has the potential to become resistant. This is useful because it means that the pathogen could be channelled into evolutionary dead-ends, stopping them from developing resistance – so preventing resistance from occurring in the first place rather than finding cures for drug-resistant infections.
