Mathematical modelling of oral microbiome resilience

Lead Research Organisation: University of Nottingham
Department Name: Sch of Biosciences

Abstract

Maintenance of oral health is essential. It is known that the microbiome, host response, and human behaviours (e.g. tooth brushing) play an essential role in maintaining oral health. It is also essential that consumer products that contact the oral microbiome do not adversely impact it. However, the oral microbiome is a highly complex system, with considerable diversity between individual people.

The aim of this project is to develop a mathematical model to characterise oral microbiome resilience, and to use the model to identify factors predictive of dysbiosis/stability of the oral microbiome1,2. Such a description would be key to determine functions of the microbiome which need to be protected to ensure consumer health as part of microbiological risk assessments. The model will be calibrated with metagenomics and metatranscriptomics data from a longitudinal clinical of experimental gingivitis currently being carried out in NJ's laboratory at the University of Newcastle in partnership with Unilever (data available Spring 2023). This is an ideal experimental system to study resilience as it is at the reversible stage of disease that may lead to irreversible periodontitis. The model will be further supported with existing data from the literature3.

Specifically, the project will:

Undertake systematic reviews of existing literature and modelling approaches to oral microbiome and measures of resilience/stability of microbiomes in general. [Year 1]

Develop dynamical mathematical models for competing microbial populations in the oral microbiome, based upon existing models developed in DS's laboratory as well as literature models. The models will include dynamics of growth (logistic or Monod terms as appropriate), death, physical removal, and competition of microbial populations. [Year 1]

Develop suitable measures of resilience/stability of the oral microbiome, which are key to the notion of dysbiosis. Preliminary work at Unilever has shown that diversity or taxa are not in of themselves good indicator(s) of health & disease. We will define metabolic functions by processing metagenome and metatranscriptome data generated at Newcastle using advanced bioinformatics analysis being developed in the Unilever internal pipeline U-discover. This analysis will identify altered metabolic pathways and functions predictive of health/dysbiosis. [Year 2]

Refine the model in the light of microbiome data, to include relevant functional populations or substrate dynamics from Objective 3. Fit the mathematical model to microbiome data from Objective 3, and appropriate literature data. These will be undertaken using Bayesian MCMC approaches, in order to infer unknown model parameters4. Priors will be taken from literature data where available. [Year 2-3]

Use local and global sensitivity analyses5 and counterfactual simulations to identify factors in the model predictive of health/dysbiosis. These will be used to inform risk analysis pipelines for Unilever [Year 3-4]

Disseminate research results, including KT within Unilever, and publication of results in peer-reviewed journals. [Years 2-4]

References (as DOIs): (1) 10.1177/0022034517742139 (2) 10.1186/s12903-019-0889-z (3) 10.1128/mBio.03281-20 (4) 10.1098/rsif.2015.0069 (5) 10.1093/femsec/fiw040

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
BB/T008369/1 30/09/2020 29/09/2028
2748357 Studentship BB/T008369/1 30/09/2022 29/09/2026