Computational Modelling

Applying Mathematical Models to Human Health Risk Assessment

In addition to experimental data from both in vivo and in vitro models and a vast collective knowledge of toxicological risks and mechanisms, toxicological risk assessors now have computational models of biological processes to bring to bear on risk assessments.

Computational models are mathematical models that use computer simulations to predict the behaviour of complex systems.  Computational models are currently in development to address the various elements of a safety risk assessment. Physiologically-based pharmacokinetic models approximate the extent of adsorption, distribution, metabolism and excretion that a chemical experiences within the human body.  Molecular modelling techniques can be used to examine the initial interaction between a chemical and the signalling and regulatory machinery of a biological pathway and thus the nature of the perturbation applied.  Systems biology models seek to model the extent to which the perturbation of biological pathways result in reversible or irreversible phenotypic changes at the cell level.

Taking advantage of new experimental and bioinformatic methods for developing and parameterising such models will provide us with the tools required to revolutionise our risk assessments.  However, mathematical models are no different from the in vivo and in vitro models traditionally used in human health risk assessment in the sense that they are an imperfect representation of reality.  Key to applying any model in human health risk assessment is to accept that we are uncertain how representative our model predictions are of reality.  Historically, the imperfections of in vivo and in vitro models were accounted for by using uncertainty factors.  Sources of uncertainty in computational models can derive from gaps in scientific knowledge, measurement and sampling errors, extrapolation, model structure and many other factors that impede our ability to make the perfect model prediction.  Understanding how we can address sources of uncertainty and account for them in our predictions from computational models of complex biological systems is a key area of research for Unilever.

Latest Presentation

Applying mechanistic modelling to human health risk assessment: a skin sensitisation case study

Harnessing mathematical models and uncertainty in toxicological risk assessments

A Bayesian approach to expert judgment of uncertainty in skin sensitisation risk assessment

Latest Publication

Gosling, J. P., Hart, A., Owen, H., Davies, M., Li, J., and MacKay, C. (2013) A Bayes Linear Approach to Weight-of-Evidence Risk Assessment for Skin Allergy. Bayesian Anal. 8, 169–186.

External links
Blog: How do we boost confidence in using mathematical modelling for toxicological safety assessments?

NC3Rs Research Grant: Uncertainty and confidence in applying mathematical models and in vitro data in toxicological safety assessments

Dr. Cameron Mackay