Data-informed modelling of rumen fermentation

Ruminant animals such as cows, sheep and goats derive their energy by converting plant matter to metabolisable nutrients. This occurs by the process of anaerobic fermentation in the rumen facilitated by a huge variety of bacteria, fungi and protozoa. The biochemistry of rumen fermentation is complex and malleable - multiple microbes are involved in the production and consumption of any particular metabolite and microbes shift metabolic niches in response to environmental changes. The complexity and malleability of rumen fermentation pose a significant obstacle when trying to predict the effect of an intervention (e.g. a change in feedstock, drug treatment) on the overall response of the system. In this project you will use a combination of experimental measurements, data analytic techniques and mathematical modelling to map the metabolic networks of rumen fermentation and their response to perturbations.

Using a rumen bioreactor you will measure metabolite concentrations and microbial community profiles and how they are affected by interventions such as antibiotics. Causal inference techniques will be used to reconstruct biochemical reaction networks from the data. You will then use bond graph models of chemical reactions to develop mathematical models of the fermentation process that are grounded in thermodynamics. The models will be validated against bioreactor measurements using inhibitors of methanogenic enzymes as well as existing field data from animal trials.

Desired skills

All candidates with a background in engineering or the natural sciences will be considered. Prior experience in measurement or modelling techniques is not required. Ability to pick up new skills and problem-solve is essential.

Find out more about this project and other projects on our website here.