Skip to Main content Skip to Navigation

Apport de la modélisation pour une meilleure stratification des populations à risque

Abstract : This PhD project was part of a system epidemiologic approach which aim to identify contributors of complex pathologies, at multiple levels, and their interactions by using system approach. This usually combine data from omics analysis and observational epidemiologic data to rich a fixed goal. The long term objective is to be able to use metabolomics to re-classify those pathologies. To this end, it is necessary to modelling phenotypes of at risk population using statistical and mathematical approaches. In this context, this project aim to characterise Metabolic Syndrome (MetS) in elderly people using multidimensional phenotyping (metabolomics, lipidomics, phenotyping, nutrion...). Different bio-informatics goals have been defined. They involve care and management of big volume of complex data, and knowledge extraction from multidimensional data in a multi-step and iterative process. From bio-informatics field, this project first allow the creation of a MetS biomarkers from literature database. It also allow the establishment of guidelines for management of data within future metabolomics projects. Thereafter, an entire and reproducible workflow for knowledge extraction have been developed, aiming to provide a signature (set of biomarkers) from multi-platforms non-targeted metabolomics/lipidomics data. It include the development of a new tool to manage analytical correlation generated on metabolomics data during acquisition and filter it ; deployment of variable selection strategies using different statistical models to provide a MetS signature for elderly people ; or execution of bio-informatics tools to go further in biological interpretation of data across visualization on metabolomics networks. Finally, a similar approach was design for sub-phenotyping of MetS modelling. It specially consist to study potential sub-phenotypes of Mets to propose a molecular re-classifying of individuals based on metabolomics data.
Document type :
Complete list of metadatas

Cited literature [403 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Thursday, October 1, 2020 - 10:45:20 AM
Last modification on : Wednesday, October 14, 2020 - 4:13:22 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02954551, version 1


Stéphanie Monnerie. Apport de la modélisation pour une meilleure stratification des populations à risque. Bio-informatique [q-bio.QM]. Université Clermont Auvergne, 2019. Français. ⟨NNT : 2019CLFAC098⟩. ⟨tel-02954551⟩



Record views


Files downloads