Foodborne Bacterial Pathogens
Microbial Complexity
Complex microbial sytems are characterised by (i) many, interacting agents; (ii) their sensitivity to small changes; (iii) emergence of properties that cannot be predicted as a sum of individual effects.
We use new mathematical modelling tools to determine the reasons behind microbial complexity. The variability of the germination time of C. botulinum is studied by stochastic probabilistic models. The adaptation of E. coli and Salmonella during lag phase is modelled by network science methods applied to the transcriptomic networks of their genes. The variability of Campylobacter responses to the environment is assessed by biostatistical techniques.
Modelling bacterial responses to the environment (predictive microbiology)
Primary Objective
- To extend the concept of predictive microbiology to include interactions, adaptive responses and other complexities.
At the level of bacterial population, and for well defined environments, we have developed models and performed validation with large datasets (see www.combase.cc). We have begun to study the randomness and uncertainty of the predictions both at the population and at the single cell levels. The sophistication of the ComBase System will be improved by including probabilistic (for low-count concentrations) and dynamic models. The database will be extended to contain records on microbial responses to dynamic environments.
Our current results are for the straightforward situation when a single cell, or a homogeneous bacterial population is in an environment defined by only a few dominant factors (such as temperature, pH) and is easily quantifiable. The modelled response is relatively simple: it reflects the growth/survival profile of the bacterial population or the distribution of a single-cell parameter (such as cell-cycle length) over the population.
However, in reality, the environment is complex and dynamic. It has many and interacting factors, for example in a complex food or in the gut, that change with time, partly due to the bacterial response. How can bacterial kinetics, and its variability, be assessed in an environment that is altered by bacterial metabolism? The "physiological state" of the cells is a concept that is routinely used but has never been quantified. Ideally, the cells' "suitability" to a given environment should be quantified. How can the transcriptomic adaptation to the environment (i.e. "non-genetic adaptation" without mutation) of a bacterial population be described by mathematical models at the molecular and system-level?
These questions will be investigated by stochastic models and network science methods. We will assess, by stochastic process modelling the effect of history and the growth environment on the distribution of single cell lag time. The physiological adaptation of bacteria to the environment will be studied by the topology of the intracellular transcriptional network.
Assessment of risks to human health associated with complex bacterial responses (Quantitative Microbial Risk Assessment)
Primary Objective
- To extend current QMRA methodologies by embracing new molecular datasets and new analyses.
Over a period of approximately ten years quantitative microbial risk assessment (QMRA) has become a central part of food safety management in both public and private domains. This progression has included agreements on accepted practice, such as the Codex Alimentarius, and initial international efforts to organise shared data and expertise; e.g. www.foodrisk.org. IFR has made substantial contributions to the development of QMRA methodology for foodborne hazards. The past decade has seen the emergent framework for QMRA to be dominated by models, or simulations, that follow the dynamics of agent populations, either cells or their associated toxins, within a prescribed food chain structure. Where possible, these kinetic models have integrated population variability and information uncertainties along the chain into representations of uncertain beliefs about recognised endpoints. Human exposures have become the most common endpoints for QMRA; this is indicative of the relative absence of information concerning human dose-response behaviours for major pathogens.
In parallel with the progression of QMRA, high-throughput 'omic technologies, have caused a revolution in the data supply relating to food safety assessment. The new information provides details, of microbial genetics and molecular microbiology for foodborne pathogens, which in many respects go far beyond the bacterial population-based data that supports the established QMRA. For many hazards the new information supply has the potential to provide additional (sub)-population quantification, to reduce uncertainties associated with established dependencies and, ultimately, to question structural (independence) assumptions for identified hazard scenarios. However the integration of specific new information sets, into system-wide representations of foodborne hazards, remains to be established.
Fax: +44(0)1603 507723 
