Small molecules in a large population: Metabolomic Profiling in the FoCuS Cohort with Data Pre-processing and Integrative Approaches

Nutritional intake, lifestyle choices and social determinants have been identified as significant contributors to interindividual metabolic variation, and play a crucial role in the development of chronic, non-communicable diseases. The present dissertation employs non-targeted metabolomics and microbiome approaches to systematically investigate these relationships in a large population-based cohort. Firstly, it is demonstrated that data merging and post-acquisition-based normalisation are crucial for statistical significance in the field of urine metabolomics. Extensive FT-ICR-MS data demonstrate that separate normalisation of individual data sets prior to merging reduces technical variability and better preserves biological signals, from which practical recommendations for population-based studies are derived. Furthermore, educational attainment as an indicator of socioeconomic inequality has been found to be correlated with metabolomic and microbial profiles. It was found that there were consistent differences in the urine and serum metabolome, as well as subtle but significant changes in the gut microbiome. Integrative analysis employing mixed graphical models has been demonstrated to link educational attainment to dietary quality, obesity and glucose metabolism markers, and selected metabolites. A third focus is the initial application of structure-based chemodiversity concepts to human urine metabolome data. The findings demonstrate that the chemical diversity of urine can effectively capture robust metabolic properties that extend beyond mere abundance. These properties are associated with a plant-based diet and gut microbiome structure, yet occur independently of individual clinical markers. It is therefore evident that chemodiversity is representative of an integrated biochemical phenotype, situated at the interface of nutrition, microbiome and host metabolism. The dissertation as a whole makes a methodological and conceptual contribution to the field of nutritional metabolomics by establishing robust analysis pipelines, identifying social and nutritional determinants of the metabolome, and introducing chemodiversity as a system-oriented analysis concept for population-based studies.