Cell systems modelling

Professor David Fell


Our work centres on modelling the networks of reactions in cells, with particular emphasis on metabolism. It forms part of the emerging field of Systems Biology, in that we are concerned with understanding how biological function arises from the interactions between many components, and with building predictive models. We have to develop and apply suitable theoretical tools, including metabolic control analysis, computer simulation and other forms of algebraic and numerical analysis. In addition, we are investigating how to decipher the metabolic information contained in genome sequences. Currently we are involved in projects on microbial, plant and animal metabolism, each in collaboration with an experimental team.

Potential applications of our work include the design of changes in cellular metabolism to improve the output of product such as antibiotics, detecting vulnerable sites in cellular networks that could be targets for drugs to control disease-causing organisms, and improved understanding of how organisms manage to adjust their metabolism in response to environmental changes and other signals.


Our work divides into the development of tools and techniques for modelling and analysis that are largely common to all our projects, and the specific application areas.

Tools and techniques:

In all our work, the first step is to create a formal description of the network that can be used for mathematical analysis, and one of our current concerns is to develop methods of model building and validation that are equal to the task of assembling networks of hundreds or thousands of reactions to represent as accurately and completely as possible the total metabolism of a cell encoded in its genome sequence.

Once we have a model, we carry out a structural analysis of the network (using techniques from linear algebra such as elementary modes analysis) to determine its capabilities; for example, determining which nutrients it can process to produce necessary cellular materials and by what routes. In some cases we can go further and build dynamic computer simulations that allow us to predict what influences the rate of metabolism. We use experimental data, firstly to provide parameter values for the models and then to test model predictions, and for these purposes we need to carry out quantitative analysis and parameter fitting. All the parts of this process are are incorporated in our own computer package for metabolic modelling: ScrumPy

Analyzing microbial metabolism:

We have been using microbial metabolism to try out our techniques of genome-scale metabolic reconstruction and analysis, starting with the well-studied Escherischia coli, where we have found that the regulation of gene transcription can be correlated with the structure of the metabolic network. The availability of large numbers of microbial genome sequences is allowing us to develop methods for comparison of different network structures. Recently we built a metabolic model of the organism used to make the antibiotic erythromycin, Saccharopolyspora erythraea, starting from its newly-determined genome sequence. We are also studying models of the metabolism of different strains of the pathogen Streptococcus agalactiae to try to understand if there are particular metabolic features related to their propensity to cause disease.

Plant metabolism:

We have a long-standing interest in photosynthesis and have applied a range of our techniques to a model of the carbon assimilation reactions in the plant chloroplast. We also recently built a computer model of potato tuber metabolism to investigate the factors that influence the extent to which they convert starch to sugars during storage (cold-sweetening). It showed that several separate changes combined to produce this undesirable outcome. Currently we are adapting the methods we have used with microbial metabolism to build and analyze a large-scale model of the metabolism of Thale cress (Arabidopsis thaliana), which has been adopted by plant scientists as their experimental model.  We are also following this up with a model of plant metabolism.

Mammalian metabolism:

We are currently part of a collaborative project to develop computer simulations of mitochondrial metabolism in mammalian cells. The mitochondria are the power houses of the cells but can develop defects that are associated with degenerative diseases and ageing. Other recent projects on the study of mammalian muscle metabolism have used metabolic control analysis as a theoretical framework to interpret how the regulatory mechanisms work.


David Fell is a consultant to Physiomics plc, an Oxford company using computer simulation to aid development of drugs and therapies, and a member of the Scientific Advisory Board of Valirx plc, a biopharmaceutical company developing innovative products.




  • Oxford Brookes University


Tools and Techniques:

  • Prof Stefan Schuster, University of Jena

Plant Metabolism:

  • Prof Piero Morandini, University of Milan
  • Dr Lee Sweetlove and the Centre for Integrative Systems Biology, University of Oxford
  • Dr Sudip Kundu, University of Calcutta

Microbial Metabolism:

  • Prof K V Venkatesh, I.I.T. Bombay
  • Dr Mark Anthony, Oxford Radcliffe Hospitals

Mammalian Metabolism:

  • Prof Jean-Pierre Mazat, University of Bordeaux 2, Victor Segalen



+44 (0) 1865 483247
See my website



Cell systems modelling Arabidopsis thaliana Calvin cycle model