Pathogen populations in their ecological and public-health context.

Working premise

The genome is not the whole story—but it is a powerful record of one.

Genome sequences retain signals of ancestry, recombination, host association, antimicrobial selection and movement between populations. Those signals become genuinely useful when they are connected to good sampling, clinical and ecological metadata, and local knowledge.

Our programme therefore moves repeatedly between scales: from nucleotide changes to lineages, from lineages to host communities, and from surveillance datasets to intervention choices.

Enteric disease is often most severe where genomic surveillance is least complete. We work with collaborators in South America, West Africa and Southeast Asia to build datasets that represent local epidemiology rather than treating established collections from Europe and North America as universal.

The objective is not simply to generate more sequences. It is to ask locally relevant questions about persistent infection, asymptomatic carriage, mixed exposure, seasonal transmission and the role of food, animals, water and household environments.

Antimicrobial resistance is produced by selection, but its distribution is shaped by population structure, recombination, mobile genetic elements and contact between ecological niches. We investigate these processes in Campylobacter, Salmonella, staphylococci and other clinically important bacteria.

A recurring focus is the movement of resistance determinants through livestock systems and food chains, including the genomic backgrounds that enable resistance to persist. This work links specific mechanisms—such as target mutations, efflux systems or plasmids—to the lineages and production systems in which they spread.

Many bacterial pathogens circulate across people, livestock, wildlife, food and environmental reservoirs. Genomic source attribution helps estimate the relative contribution of those reservoirs to human disease, while carefully designed field studies reveal the pathways through which exposure occurs.

We develop and evaluate machine-learning and probabilistic models, paying particular attention to population structure, uneven sampling and the need to quantify uncertainty. The purpose is not a single permanent percentage, but a surveillance framework that can detect changes and support targeted control.

Pathogens can cause acute disease, prolonged carriage, recurrent infection or no obvious symptoms. We use genomic and epidemiological data to investigate whether these outcomes reflect bacterial lineage, host context, exposure intensity, mixed infection or interactions with the wider microbiome.

Current interests include persistent Campylobacter infection and childhood growth, lineage-specific disease associations, severe or unusual presentations, and the possibility of identifying genomic markers that improve risk stratification.