Research
Building bespoke, predictive, mathematical models of biological systems.
I enjoy building bespoke mathematical models of biological systems. Deciding which mechanisms to include in a model is a bit of a dark art. One simplifies the biological system to within an inch of its life, adds assumptions in place of unknown or complex data, approximates parameters, etc, yet still these simple models have the power to predict the outcome of yet unperformed experiments, override misconceptions, and give us a deep understanding of mechanism underpinning behavior.
I have spent a lot of time building uninformative models. But I have also been lucky enough to work with some great scientists in building models that have changed our understanding of a system. For example:
1) Savage,N.S., et al. (2008). A Mutual Support Mechanism through Intercellular Movement of CAPRICE and GLABRA3 Can Pattern the Arabidopsis Root Epidermis. PLoS Biology,6(9). doi:10.1371/journal.pbio.0060235.
In Savage 2008 I developed a unique mathematical framework designed to minimize modelling assumptions and take full advantage of on/off biological data. The model predicted that the current theory of WERWOLF self-promotion, in the Arabidopsis root epidermis, was incorrect and proposed an alternative hypothesis. The alternative hypothesis, generated by the model, was verified experimentally within the publication. The following year Kang 2009 published further experimental verification showing that WERWOLF did not regulate its own expression.
2) Savage,N., et al. (2013). Positional signaling and expression of ENHANCER OF TRY AND CPC1 are tuned to increase root hair density in response to phosphate deficiency in Arabidopsis thaliana. PloS one,8(10). doi:10.1371/journal.pone.0075452.
To ascertain the mechanisms by which phosphate may be influencing root hair numbers in Arabidopsis I used probability theory, as there was no knowledge of possible mechanism at that time. Using the analysis of root epidermal cell numbers, cell lengths and cell fate specification, in plants grown under various phosphate conditions, we hypothesized that in phosphate deficient conditions there are two mechanisms working together to increase root hairs per unit area. In 2018 both hypothesized mechanisms were supported experimentally (Rishmawi 2018).
3) Savage,N.S., et al. (2012). Mechanistic mathematical model of polarity in yeast. Molecular Biology of the Cell, 23(10). doi:10.1091/mbc.E11-10-0837.
4) Dyer,J.M., Savage, N. S., et al. (2013). Tracking Shallow Chemical Gradients by Actin-Driven Wandering of the Polarization Site. Current Biology,23(1). doi:10.1016/j.cub.2012.11.014.
In Savage 2012 we developed a unique model coupling reaction diffusion equations, describing cell polarity on the yeast plasma membrane, with the computational physical addition of plasma membrane to the cell surface, describing cell growth. This work hypothesized that a previously published and universally accepted hypothesis was incorrect, that the formation of a polarity patch on the yeast plasma membrane was not the result of two positive feedback loops, but rather one positive and one negative. In Dyer 2013 the group generated experimental data to show that 1) the negative feedback loop proposed in Savage 2012 existed in yeast, and 2) it was an essential mechanism for gradient tracking during mating.
Developing bespoke mathematical frameworks.
In order to build a model that uses the available data to its full capacity, and addresses the question at hand, on occasion, one must merge theory from disjoint areas.
My interest in building innovative models started during my PhD after reading Kauffman 1969. Here I merged Random Boolean Networks with Markov theory to generate an information network with prescribed 'accuracy of information'. I developed this to help British Telecom address their bandwidth capacity and pricing issues - obviously fiberoptics made the work was obsolete before it was completed.
Papers 1 and 3 discussed in the above box, and the publications, Savage 2020, Ishitsuka 2015, and Okoda 2013, are further examples of unique modelling frameworks, designed specifically for the data and question being addressed.
The molecular control of hyphal morphology.
Here we address the specific question of the molecular control of hyphal cell morphology.
This work combines wet and dry methodologies and was funded by the BBSRC.
Research grants
Specifying patterns in 3D: Epidermal patterning of roots
LEVERHULME TRUST (UK)
December 2021 - January 2025
Control of Hyphal Gowth
BIOTECHNOLOGY & BIOLOGICAL SCIENCE RESEARCH COUNCIL
July 2016 - June 2019
YEAR 1 Wellcome Trust ISSF Non-Clinical Fellowships
WELLCOME TRUST (UK)
October 2011 - May 2018
Modelling gradient sensing and collective cell migration in the Drosophila ovary
BRITISH SOCIETY FOR CELL BIOLOGY (UK)
June 2014 - August 2014