Metabolic profiling has been revolutionized during the last 'post-genomic' decade: several aspects of analytical chemistry, statistics applied to chemistry (i.e. chemometrics), and metabolic biochemistry have merged together to form a new and very dynamic, truly interdisciplinary area of science. Among the scientific areas covered by metabolic profiling methods, metabonomics and metabolomics have independently emerged from distinct disciplinary fields. As a result, the field of metabonomics/metabolomics was ranked fourth out of '10 Emerging technologies' listed in MIT's Technology review in 2005. Biological samples are analyzed by NMR and the spectra are imported in a database. Then the spectral information is summarized in multivariate statistical models such as principal component analysis, linear discriminant analysis or partial least square regressions to derive significant biomarkers associated to the condition under scrutiny. The power of the metabonomic approach lies in the fact that there is no a priori expectation in metabolite level variation, which guarantees a strong potential for biomarker discovery. The bottleneck for advancements in the field of NMR – based metabonomics is twofold: (i) the rarity of metabolic profiling dedicated NMR pulse sequence and (ii) the statistical analysis of NMR data themselves. The objective of this proposal is to develop a metabonomic platform with a new generation of NMR and chemometric tools needed to analyze functional metabolic network data that can be used in functional genomics of model organisms, disease prediction in clinical studies, and therapeutic target discovery in drug development. We have identified specific components of the approach that require development with the following main objectives within: (i) development and optimisation of metabolic profiling oriented NMR pulse programs for analysis of biopsy and biofluid samples, (ii) generation of a pure compound standard database including biologically relevant meta-data (in collaboration with Bruker GmbH), (iii) implementation of state of the art chemometric tools for automated metabolite quantification and pattern recognition models for network and systems biology. We intend to achieve these objectives within the context of the following highly novel biological applications: (i) functional genomics in Caenorhabditis elegans, (ii) screening inborn errors of metabolism in human foetal urine, (iii) diagnosis / prognosis of cancer. The results are expected as follows: development of NMR pulse sequences and chemometric regression models inferring metabolite concentrations in 1H NMR mixture spectra from buifluids and biopsies, using a pure compound standard library, i.e. a spectral dictionary, in collaboration with Bruker GmbH. Different deconvolution and curve resolution approaches will be compared ie, PLS vs. bayesian models (collaboration with University of Oxford). In this context, we expect the increase in spectrometer field strength (initially from 600MHz to 700 MHz then 900 MHz and 1 GHz) to result in: (i) better metabolite concentration inference from 1H NMR mixture spectra, better disease prediction in animal models or human cohorts. Eventually, we will apply network biology visualisation tools based on graph theory. Once reliable concentrations are obtained from NMR spectra, we expect network biology to become the new standard for representing metabonomic data. On the application side, we target: (i) a proof of concept on the use of metabonomics to reveal silent mutations in Caenorhabditis elegans, which would represent a major breakthrough for the C. elegans functional genomics community, because NMR-based metabonomics would provide a cost-effective tool for discovering gene functions, (ii) a high throughput foetal urine screening programme by NMR to provide paediatricians with mechanistic information about disease not routinely available, (iii) a pharmaco-metabonomic study of the effects of growth hormone on cancer cells to develop the potential of very high field NMR for diagnosis and mechanistic purposes.