| Summary: | Invasive Staphylococcus aureus disease is an important cause of morbidity and mortality, but is much rarer than asymptomatic carriage. The contribution of the bacterial genome to S. aureus infection is incompletely understood, and molecular epidemiology provides conflicting evidence. This thesis aims to improve our understanding of this contribution using the resolution afforded by whole-genome sequencing. In a systematic study of S. aureus evolution in 105 hosts during invasive S. aureus disease, I demonstrate extensive within-host diversity, with evidence for varying selective pressures and within-host adaptation. Evidence for adaptation is strongest in genes under control of transcriptional regulatory systems, including Repressor of surface proteins (Rsp) (p=10<sup>-6.4</sup>) - a recently discovered global virulence regulator - and the Accessory gene regulator (Agr) (p=10<sup>-5.6</sup>), which are enriched for protein-altering variants 3.6- and 2.9-fold respectively. The development of invasive disease is associated with subtle changes in the transcriptional regulation of Staphylococcus aureus arising within hosts. Applying recently developed tools for bacterial genome-wide association studies (GWAS), I present GWAS investigating for S. aureus genomic associations with two disease phenotypes: bacteraemia and pyomyositis. A study of S. aureus bacteraemia and carriage in 2001 isolates from the United Kingdom shows bacteraemia is not strongly bacterially determined: no lineages, genes or variants were significantly associated with bacteraemia. In a study of 518 isolates from pyomyositis and carriage in Cambodian children, the presence of Panton-Valentine leukocidin (PVL) genes increases the odds of pyomyositis 130-fold (p=10<sup>-18</sup>), and variation in these genes and an adjacent promoter region are sufficient to explain over 99.9% of the heritability of pyomyositis. These results establish staphylococcal pyomyositis, like tetanus and diphtheria, as a disease depending critically on expression of a toxin. Microbial genomics offers unparalleled opportunities to understand infections, and here I demonstrate insights generated through pathogen evolution within hosts and bacterial GWAS.
|
|---|
