Statistical Methods for Whole Genome Phenotype Mapping in Bacterial Populations. (360G-Wellcome-101237_A_13_Z)

Pathogenic bacteria cause severe disease worldwide. In Britain, hospital- associated infections of Staphylococcus aureus and Clostridium difficile are major causes of death. Strains differ in virulence, but the genetic variants responsible are largely unknown. Breakthroughs in sequencing provide new opportunities to map virulence, but the lack of analysis tools represents a roadblock to discovery. My goal is to develop an analysis framework for whole- genome phenotype mapping in natural bacteri al populations. Focusing on thousands of hospital-associated pathogen genomes, I will 1. Investigate the influence of common genetic variants on bacterial virulence. 2. Investigate the role of de novo mutation in bacterial pathogenesis. 3. Optimize analysis tools for bacterial population structure and linkage disequilibrium. This project fuses evolutionary genetics, computational statistics, clinical microbiology and public health, pursuing work I led into virulence evolution during coloni zation. Building on genome-wide association studies, it proposes solutions to the challenges of analysing bacterial populations arising from clonal reproduction and horizontal gene transfer. This project could help identify variants/pathways associated with virulence that lead to new drugs and assays for improving patient treatment and prioritizing public health resources. The tools will help scientists mapping diverse bacterial phenotypes relevant to health, agriculture, ecosystems and the e conomy.