Abstract

Understanding the within-host dynamics of acute bacterial infections is vital for the development of optimised treatment and the management of drug resistance. A major hurdle has been the limited resolution of data that can be obtained during infection, especially when dealing with low numbers of bacteria. For the last 10 years, a multi-disciplinary research programme at the Department of Veterinary Medicine has developed a pioneering approach, combining “wild-type isogenic tagged strains” (WITS) of Salmonella enterica in mice with mathematical models, to quantitate the key processes that drive the early stages of infection. In this talk, I will present the latest analytical methods that our group has been exploring in order to improve statistical inference and, ultimately, help optimise experimental design. I will cover likelihood-based and likelihood-free techniques, and highlight an efficient moment-closure approach which is showing great promise.