Advances in bioinformatics and genome research have generated a rapid expansion in the availability of information at all levels of biological investigation. A computational infrastructure for systems biology is needed to interpret this information at the whole cell level, and to predict behavior of complex systems in response to their environment. This project will develop a combined experimental/in silico platform to improve the predictive capabilities of models using metabolite profile (metabolomics) data. Specifically, within existing metabolic modeling software, an infrastructure will be developed to manage, visualize, and analyze metabolomics data in the context of the genome-scale model. In Phase I, qualitative metabolomics data was used to improve the quality of existing genome-scale models, by finding gaps in the network and identifying candidate genes with functions to fill these gaps. Next, quantitative concentration data was used in conjunction with thermodynamic considerations to probe intracellular metabolism and improve the ability to predict cell physiology. Two computational methods were implemented and tested with available datasets to predict reaction directionalities, identify potential bottleneck sites, and predict potential sites for regulation. Then, the capabilities of performing such integrated analysis will be demonstrated, using engineered E. coli strains as a case study.
Commercial Applications and Other Benefits as described by the awardee: A general methodology for extracting useful biological information from metabolomics data not only should increase modeling capabilities but also should guide rational strain engineering for the production of chemicals and fuels from renewable feedstocks
