Natural microbial communities perform many vital engineering functions in wastewater treatment, bioenergy production, and bioremediation, but our understanding of how these complex communities assemble, function and respond to environmental change is limited. In the past we have used an empirical approach to utilising these systems based on accumulated knowledge. This has had some success, for example in wastewater treatment, but previous experience will not help us deal with novel environments and problems. To start engineering these systems and optimise them for different applications we need mathematical models that are capable of predicting their behaviour. The urgent need to harness microbes to their full potential has been brought into sharp focus by the climate, energy, pollution, and water crises we now face. Development of predicative models has been constrained by the difficulty of obtaining information on these communities. Historically analysis was restricted to those microbes that could be isolated and grown in the laboratory but these represent only a fraction of the community. The new science of environmental genomics, by direct extraction and amplification of DNA, has sidestepped the need to culture organisms, but until the last few years actual sequencing of this DNA was slow and expensive. Consequently sample sizes were small compared to the huge diversity and numbers of microbes. Now new high throughput sequencing technologies are available, which have increased the rate of data acquisition by orders of magnitude, allowing us for the first time to obtain a detailed picture of the composition of these communities and how they vary through space and time. Using metagenomics we can also start linking the identity of the community members to their metabolic functions. Finally, we have sufficient data to start constructing the models we desperately need. This fellowship will exploit this opportunity, through an integrated approach, to develop a new combined genomics modelling paradigm for the study of microbial systems. New statistical tools and software will be developed to filter noise from the sequencing data, and extract information which can then be fed into multi-scale mathematical models. At the most fundamental level these models will have an explicit description of individuals moving, reproducing and interacting through the consumption and production of chemical substrates. Using advanced mathematical techniques they will be scaled-up to a description of whole populations. This will enable us to define the models in terms of processes operating on the level of individuals but validate them with the genomics data which provides a population level picture. It will also allow these models to be applied on the whole system scales necessary for their industrial application. The statistical tools, and mathematical models developed will be completely generic but we will illustrate the approach by focusing on two specific case studies: low temperature anaerobic wastewater treatment and microbial fuel cells. The former has the potential to reduce the energetic costs and carbon footprint of the treatment of wastewater in the UK and other temperate countries. The latter could provide a cheap, clean source of electricity in remote locations using virtually any organic substance as fuel. They could be particularly useful in the developing world. We will explore further applications of our paradigm to other vital microbial based engineering problems. In addition, the tools and models developed will be applicable to the study of microbial communities in any area, from human health to the biogeochemical cycles that sustain life on this planet.