PhD Studentship: Machine Learning for plasmid risk assessment from genomic surveillance

Respiratory infections are a major cause of morbidity and mortality, however, important gaps remain in our understanding of their aetiology, early risk stratification, and the role of microbial communities in disease.

This project will apply advanced computational methods to analyse large-scale metagenomic sequencing data from patients across the respiratory disease severity spectrum, ranging from adults presenting with respiratory symptoms in A&E to patients with severe pneumonia in intensive care. The project aims to define how the respiratory microbiome changes with disease severity.

Initial work will focus on microbial profiling and genome reconstruction to characterise bacterial, viral, and fungal communities. Particular attention will be given to the virome, including bacteriophages and anelloviruses, as potential indicators of microbial imbalance and disease progression. The project will investigate interactions between microbes, with a focus on phage–bacteria dynamics, and define ecological states associated with stable and disrupted communities. Machine learning models will then be developed to integrate bacterial, viral, and ecological microbiome features for improved pathogen detection, severity classification, and early identification of patients at risk of deterioration. The project will deliver both fundamental insights into respiratory microbial ecology and practical computational tools with potential clinical application.

The aims of this project include:

• To characterise the respiratory microbiome across a spectrum of disease severity and antibiotics exposure, using metagenomic sequencing data.
• To investigate interactions between bacteria, viruses including bacteriophages, and other microbes, and define ecological states of infection.
• To develop and validate predictive models for pathogen detection, disease severity classification, and risk of clinical deterioration.
• To assess the role of the virome as a complementary source of diagnostic and prognostic information.

This project requires a computational/bioinformatics background and will provide exposure to clinical and public health research. It offers opportunities to collaborate with clinicians and microbiologists, supporting the development of both technical and translational skills.