Statistical methods for intake prediction and biological significance analysis in nutrimetabolomic studies

Abstract

[eng] This thesis is the product of three and a half years working on the complex world of metabolomics and nutrition. All the work presented here is focussed on the problems arising from associating and integrating metabolomics data with nutritional or dietary data. This issue has been approached using both observational and interventional studies and from a mainly bioinformatic point of view, proposing different methods and tools to reduce the complexity of nutrimetabolomics data analysis. Thus, this work consists of four chapters divided into three parts, in addition to a summary of the content of the entire thesis in Catalan, the references, and the appendices. The first part consists of a global introduction, where the fundamental concepts needed for the correct understanding of the thesis are reviewed, as well as basic concepts about metabolomics and nutrition, the state of the art of the nutrimetabolomics field, and the fundamentals of biological significance analyses, among others. Then, this first part ends with a brief definition of the objectives of this work. In the second part, the results of this thesis are carefully presented and discussed. The results are presented in a compact format, with each section being a summary of a scientific publication. These results include the develompent of an ontology that defines the relationships between dietary metabolites and foods, the development of an open source tool for metabolomics data analysis, the development of an open source tool for nutrimetabolomics enrichment analysis, other open source tools developed in the context of this work, and a section with different publications where the methods and tools developed have been applied. Then, all these individual results are discussed together, providing a global and unified context where all the developments of this thesis are related. Lastly, the third part of this thesis presents the conclusions, contextualizing all the obtained results within the main objective of the thesis: contribute to the improvement of the integration and interpretation of nutrimetabolomics data. Additionally, in the appendices, the published results and some extra information used in carrying out this research are presented. Finally, although this thesis is made up of contents from the fields of metabolomics, nutrition, bioinformatics and biostatistics, it has been written for a wide scientific audience, trying to be as comprehensible as possible for any profile of researchers, avoiding unnecessary complexities and always following the transversal objective of the thesis. I hope you find it useful but, above all, that you enjoy reading it.