CONGRATULATIONS TO OUR NEW DOCTOR, Christophe BECAVIN, who graduated with the Very honorable grade !
Christophe BECAVIN presented his PhD defense on December 2h 2010 at the Paris Descartes University, Faculty of medicine Amphithéâtre Lutton.
Thesis title : ""Dimensionality reduction and pathway network analysis of transcriptomic data: Application to T-cell characterization""
Members of the jury :
Rapporteurs: Alain Arneodo, Laboratoire Joliot-Curie, ENS Lyon
Lars Rogge, Institut Pasteur, Paris
Examinateurs: Jean-Marc Victor, LPTMC, Paris 6
Sylviane Pied, Institut Pasteur, Lille
Andrei Zinovyev, Institut Curie, Paris
Thesis supervisor : Arndt Benecke, IHES, Paris
Abstract :
In the context of whole-genome expression (transcriptome) data analysis, different tools already exist today. One class of tools, called
dimensionality reduction techniques, seeks for general patterns and important components of the system which can help to summarize the data.
During my thesis I extensively studied the different state-of-the-art techniques existing in this field. We then developed our own approach based on the combination of Singular Value Decomposition and Multidimensional Scaling. We proved its usefulness and accuracy.
In addition to gene expression-specific data analysis tools, we developed a software which allows to map different gene expression patterns to protein-protein networks. In order to link the gene expression scale to the protein scale (proteome). Those protein-protein networks are built based on curated ontology-based pathway models.
The tools developed here and many others were used in order to analyze different "omics" data. The first application was on the analysis of experiments measuring autoantibodies and cytokine expression in the human
body during Malaria infection. We determined specific markers of Cerebral Malaria, which will help to better detect the disease. The larger analysis we have performed, consisted in defining the transcriptome profile of regulatory T-cell subsets (Treg). These cells are depleted during HIV infection, for this reason a good molecular characterization of the different subsets would help find more accurate markers to, for example, follow their evolution during the treatment with novel drugs to fight AIDS. Among the new molecular markers of Treg we identified, a newtranscription factor FOXLF was discovered which may play an important role in the regulation of the "regulatory" function of those cells.