Process & Systems Modelling

Modelling the production of metabolites by colonic bacteria

The large intestine contains a complex bacterial community which is responsible for production of metabolites that play an important role in colonic health. Competition of bacterial species for nutrients from the digesta is poorly understood, but data are becoming available from experimental studies carried out at RINH. We have developed a deterministic dynamical model of such experimental systems which includes parameters for bacterial characteristics such as metabolite production rate. Bayesian methods have been implemented to infer bacterial characteristics, combining prior knowledge of the underlying microbial relationships with data from multiple experiments. Our approach enables prediction of bacterial communities in situations that are not easily studied experimentally, whilst properly accounting for uncertainties in estimated microbial characteristics. Simulation studies have shown our approach to work better than computationally more straightforward approaches.

starch diagram and simulation study curves For a simple experimental set up where Bacteroides (Bac.) feeds on starch to produce acetate (left), the results of a simulation study (right) show our preferred method of combining data sets across experiments is approximately unbiased (red curve) whilst a computationally simpler alternative is biased (green curve). The value of the acetate production rate used to simulate the data is given by a vertical line (white).

Further details from: Grietje Holtrop

Article date 2011


Statistical Genomics and Bioinformatics

Process and Systems Modelling

Statistical Methodology

PhD Opportunities

Meetings & Seminars