ED Mathématiques et Informatique
VIANA: Visual analytics for tire design
by Vasile CIORNA (LaBRI - Laboratoire Bordelais de Recherche en Informatique)
The defense will take place at 13h00 - 136 5, Avenue des Hauts Fourneaux, L-4362 Esch-sur-Alzette, Luxembourg
in front of the jury composed of
- Guy MELANÇON - Professeur des universités - Université de Bordeaux - CoDirecteur de these
- Mohammad GHONIEM - Directeur de recherche - Luxembourg Institute of Science and Technology - CoDirecteur de these
- Pascal BOUVRY - Professeur des universités - Université du Luxembourg - Examinateur
- Guillaume BLIN - Professeur des universités - Université de Bordeaux - Examinateur
- Alexandru TELEA - Professeur des universités - Utrecht University - Rapporteur
- Silvia MIKSCH - Professeure des universités - Vienna University of Technology - Rapporteur
With the acceleration of technical breakthroughs in the past decades, more and more technical means are available to support companies in their product developments. While not new, lately, there is a clear expansion in the utilization of machine learning models within the internal processes of companies. However, the way these models are utilized is influencing the risk-reward balance as well as their adoption significantly. This project reports reflections on the design of visualizations to support effectively the utilization of machine learning methods in the frame of product design. We identify model explainability and rapid what-if scenario creation as central pillars for utilization and adoption. These learnings were integrated into two new visualization applications. Firstly, Interact is a novel What-If tool for regression models and an explainable artificial intelligence method. The tool provides single-model support to test and understand a model, validate its mechanisms through the utilization of statistical interactions, and anticipate on actions. The evaluation of this tool with experts from the tire industry revealed that the design of the user interface widgets was key for its effective utilization. More precisely, feedforward information and navigation cues were appreciated by our evaluators. Interact includes simple widgets that convey such information. However, our users requested, in addition, new capabilities for multi-model support. This new direction called for a more in-depth analysis of prior work on how to design and use navigation cues and embedded visualizations to convey feedforward information. Accordingly, we surveyed the visualization literature for work including “scented” or sophisticated widgets. An online browser allowing exploration of relevant work comprising “scented widgets” is made available. This literature survey allowed to identify a research opportunity related to suitable visualizations for multi-models or tradeoff contexts. Respectively, we conducted a crowdsourcing experiment to compare line charts and colorfields as effective cues for multi-inputs, multi-outputs contexts. Line charts showed better performance compared to colorfields for novice users. Secondly, VisDM, builds on previous experiences such as the development of Interact and focuses on multi-inputs, multi-outputs with a specific emphasis on decision-making in product design. The tool allows using simultaneously up to 7 models that work together to assess tradeoffs. Scented widgets allow anticipating on actions effectively, and Multi Criteria Decision Making methods allow deciding on competing alternatives.
Optimal parameters determination for the execution of MPI applications on parallel architectures
by Richard SARTORI (LaBRI - Laboratoire Bordelais de Recherche en Informatique)
The defense will take place at 9h00 - TBD 1 avenue du Dr Albert Schweitzer, 33400 Talence
in front of the jury composed of
- Guillaume MERCIER - Maître de conférences - Université de Bordeaux - Directeur de these
- Soraya ZERTAL - Professeur des universités - Université Paris-Saclay - Rapporteur
- Philippe CLAUSS - Professeur des universités - Université de Strasbourg - Rapporteur
- Patrick CARRIBAULT - Directeur de recherche - CEA LiHPC - Examinateur
- Emmanuelle SAILLARD - Chargée de recherche - INRIA BORDEAUX SUD-OUEST - Examinateur
- Olivier BEAUMONT - Directeur de recherche - INRIA BORDEAUX SUD-OUEST - Examinateur
Supercomputers are used to tackle complex numerical problems, such as simulations, weather forecasting, or artificial intelligence, that require significant computational resources, inaccessible to traditional computers. Composed of multiple powerful interconnected computers, their capabilities continue to grow. However, developing applications that can fully leverage this power is becoming increasingly difficult. Indeed, various factors must be taken into account: heterogeneous computing units requiring specific programming methods, memory hierarchy, data transfers, network communications, and task scheduling. To overcome these challenges, the MPI standard was created, providing a unified interface to facilitate the programming of supercomputers and the management of communications between their components. In this thesis, we explore various ways to improve the performance of applications using MPI by adjusting various parameters to make the best use of the hardware resources available on supercomputers. The work focuses on Open MPI, an open-source implementation of MPI, and proposes optimization techniques to reduce the time and resources required for this task. Among these, we explore Black Box Optimization (BBO), an application-agnostic method that allows for efficient exploration of the parameter space. The main idea behind BBO is to use smart heuristics to select the most promising points in the parameter space to evaluate. Instead of testing all possible combinations, the BBO method enables exploration of this space while minimizing the number of necessary evaluations. Compared to exhaustive exploration, the time required to conduct the process is significantly reduced while providing a solution with virtually identical performance. This approach is particularly useful in the context of MPI applications, where the parameter space can be enormous. On the other hand, we also introduce the concept of skeletonization of MPI applications, a new technique that accelerates the optimization process by extracting a "skeleton" of the application. The extraction is automated using tools provided by LLVM, a set of tools operating around an intermediate representation. This makes the skeleton extraction independent of the programming language used for the application. This skeleton preserves the essential elements for optimizing the application, including its communication patterns, while eliminating computational aspects. It can then be used in place of the original application in the optimization process, reducing the time required for execution without compromising the validity of the obtained results. The contributions presented demonstrate that fine-tuning the parameters of the MPI implementation has the potential to improve the performance of HPC applications.
Asymptotic preserving numerical methods for moment models of plasmas physics
by Thomas VIGIER (IMB - Institut de Mathématiques de Bordeaux)
The defense will take place at 10h00 - Salle de conférence de l'IMB Institut de Mathématiques de Bordeaux Université de Bordeaux 351, cours de la Libération - F 33405 TALENCE
in front of the jury composed of
- Luc MIEUSSENS - Professeur des universités - Université de Bordeaux - Directeur de these
- Bruno DESPRéS - Professeur des universités - Laboratoire Jacques Louis Lions - Rapporteur
- Thomas REY - Professeur des universités - Laboratoire J.A. Dieudonné - Rapporteur
- Julien MATHIAUD - Professeur des universités - Institut de recherche mathématique de Rennes - CoDirecteur de these
- Teddy PICHARD - Professeur assistant - CMAP, Centre de Mathématiques Appliquées - Examinateur
- Rodolphe TURPAULT - Professeure des universités - Université de Bordeaux - Examinateur
- Héloïse BEAUGENDRE - Professeure des universités - Université de Bordeaux - Examinateur
- Anaïs CRESTETTO - Maîtresse de conférences - Laboratoire de Mathématiques Jean Leray - Examinateur
As part of the research on energy production through inertial confinement fusion, the development of new numerical methods is crucial. Indeed, simulations not only allow to refine the understanding of complex physical phenomena occurring under the extreme thermodynamic conditions necessary for fusion but also permit to assist in the design of new experimental devices. Within numerical simulations for fusion, the calculation of the electron heat flux is a central issue. Due to the extreme conditions of matter, the conventional Spitzer-Härm theory becomes insufficient to describe thermal conduction. Thus, to capture kinetic effects at a lower computational cost, moment models at the mesoscopic scale are used instead to describe electron behaviour. This thesis manuscript focuses on the numerical resolution of these moment models, first in the simplified framework of linear transport and then in the context of electronic transport. The multi-scale nature of these models complicates the development of numerical schemes, which must accurately resolve all regimes: to do this, a particular class of methods, known as asymptotic-preserving methods, has emerged. One of the most recent and promising one is the Unified Gas Kinetic Scheme (UGKS): this finite volume scheme for relaxation-type kinetic equations relies on the integral solution obtained from the method of characteristics. In this manuscript, a method is proposed to derive asymptotic-preserving kinetic schemes for moment models based on UGKS. The main idea which is introduced is to apply a closure at the numerical scale in the UGKS fluxes. In the two studied application cases, this new method reveals to be particularly effective and continuing its development seems relevant in order to solve more complex physical models. Moreover, its generic nature and flexibility make it a pertinent alternative to conventional asymptotic-preserving schemes. Moreover, some stability results are also demonstrated for this new scheme, and an extension to unstructured meshes is proposed.
ED Sciences Chimiques
Bioinspired interface engineering of DGL-based hydrogel with high mechanical performance and bioactive properties for in vitro biomaterial platform.
by Michele VALEO (Institut de Chimie & de Biologie des Membranes & des Nano-objets)
The defense will take place at 14h00 - Salle Manon Cormier (rez-de-jardin) Salle Manon Cormier (rez-de-jardin) avenue Léon Duguit, bat. G campus Montesquieu 33600 Pessac
in front of the jury composed of
- Adeline GAND - Maîtresse de conférences - CY Cergy Paris Université - Examinateur
- Etienne DAGUE - Directeur de recherche - LAAS-CNRS, Université de Toulouse - Rapporteur
- Christophe EGLES - Professeur - UMR CNRS6270 Université de Rouen Normandie - Rapporteur
- Michael MOLINARI - Professeur - CNRS UMR5248, CBMN, Université de Bordeaux - Examinateur
- Marie-Christine DURRIEU - Directrice de recherche - CNRS UMR5248, CBMN, INSERM, Université de Bordeaux - Directeur de these
Bone grafting remains the standard of care for non-union fractures or reconstructing large bone defects. Recently, tissue engineering approaches have aimed to replace bone grafting by implanting engineered biomaterials that stimulate the differentiation of endogenous bone precursors, such as mesenchymal stem cells (MSCs). MSCs are multipotent stem cells found in the bone marrow, capable of differentiating into all the tissue types found in a skeletal segment: osteoblasts (bone), chondrocytes (cartilage) and adipocytes (fat). Owing to this regenerative potential, MSCs have been extensively investigated in bone tissue engineering. The specific properties of the stem cell tissue microenvironment - the niche - are crucial regulators of their differentiation fate. Besides biological factors, MSCs can respond to specific physical cues, such as the stiffness and topography of the extracellular matrix (ECM). They were shown to differentiate into bone cells when adhering to materials that mimic the stiffness of bone ECM. Therefore, biomaterials for bone tissue engineering must present specific mechanical, topographical, and biochemical properties to direct MSC differentiation towards a desired lineage, like bone. However, because of the inherent complexity of in vivo niches, predicting the biological interactions of stem cells with biomaterials remains challenging. Due to their structural similarity with the ECM, hydrogels have emerged as biomimetic culture substrates to recreate in vitro cell microenvironments with tailorable biophysical and biochemical properties. Hydrogels can be engineered to recapitulate the essential features of the niche, allowing the study of stem cell behaviour in response to selected environmental clues. Among the various chemical strategies of hydrogel synthesis, multivalent crosslinking macromolecules, like dendritic polymers, have remained poorly explored. Exhibiting unparalleled surface multivalency, dendritic polymers can react with multiple polymer chains and form hydrogel networks with variable crosslinking density. Dendrigrafts of poly-L-lysine (DGLs) are amine-terminated dendritic macromolecules composed of lysine units. Already investigated for hydrogel crosslinking, DGLs are available in four Generations (G2-G5), are biodegradable and water-soluble. Containing approximately a thousand lysine units, DGLs of Generation Five (DGL G5) have never been explored for hydrogel synthesis. Hence, this PhD project focused on developing DGL G5-based biomimetic hydrogels with controllable stiffness and surface bioactivity as an in vitro 2D-culture system. The hydrogels were designed to improve the adhesion and differentiation of human mesenchymal stem cells (hMSCs) towards the bone lineage. Hydrogels were formed by amide crosslinks between DGL G5 and the carboxylic groups of synthetic (a modified poly(ethylene) glycol) or natural (hyaluronic acid, tested preliminarily) polymers. Hydrogel stiffness, tuned between 8 and 100 kPa by adjusting the concentration and the stoichiometry of reactants, was measured at multiple length scales using unconfined compression, oscillatory rheometry and atomic force microscopy. To allow hMSC attachment and differentiation, the hydrogels were functionalised with an equimolar solution of an RGD-containing ligand and an osteoinductive factor, a biomimetic peptide derived from the bone morphogenetic protein 2. The functionalised hydrogels improved early hMSCs' adhesion, with cells exhibiting greater spreading after 6h compared to glass controls with serum. After two weeks, hMSCs seeded on DGL-based hydrogels exhibited higher expression levels of late osteogenic markers, particularly on 22 and 30 kPa-PEG gels. These results suggest that DGL-based hydrogels can be engineered with superior bioactivity to traditional culture substrates and improve adhesion, spreading and differentiation of hMSCs.
Anionic order/disorder within Mg and transition metals (Ni, Cu, Zn) hydroxyfluorides :correlation between composition / structural features and physical-chemical properties
by Helies BOUMALI (ICMCB - Institut de Chimie de la Matière Condensée de Bordeaux)
The defense will take place at 10h00 - Amphithéâtre ICMCB – UMR5026 87, Avenue du Docteur Schweitzer 33608 PESSAC cedex (France) Suivez-nous sur Twitter
in front of the jury composed of
- Alain DEMOURGUES - Directeur de recherche - CNRS - UMR 5026 - Directeur de these
- Laurence CROGUENNEC - Directrice de recherche - ICMCB - Examinateur
- Fabrice LEROUX - Directeur de recherche - Institut de chimie de Clermont-Férrand - Rapporteur
- Vincent MAISONNEUVE - Professeur - Université du Mans - Institut des Molécules et Matériaux du Mans - Rapporteur
- Fabien GRASSET - Directeur de recherche - NIMS / Saint-Gobain / CNRS - Examinateur
In transition metal based hydroxyfluorides, fluoride (F-) and hydroxide (OH-) ligands coordinating metals are coexisting in the structure. The large difference of electronegativity/polarizability of these anions permit a tunability of chemical bonds, giving rise to a fine control on the structural features (coordination polyhedra, hydrogen bondings…) and consequently on the chemical and physical properties such as acido-basicity, thermal stability, or redox properties. The subject of this thesis concerned the preparation and the characterization of these new structures. Long-range and short-range order characterizations were performed in order to understand better the impact of the F/OH ratio on the structural features and the physico-chemical properties
Tailored syntheses of dedicated multi-scale porous materials toward environmental remediations
by Antoine VARDON (Centre de Recherche Paul Pascal)
The defense will take place at 10h00 - Amphithéatre Centre de Recherche Paul Pascal, 115 Avenue du Dr Albert Schweitzer, 33600 Pessac
in front of the jury composed of
- Rénal BACKOV - Professeur des universités - Université de Bordeaux - Directeur de these
- Marco DATURI - Professeur des universités - Université de Caen Normandie - Rapporteur
- Nicolas KELLER - Directeur de recherche - Université de Strasbourg - Rapporteur
- Thierry PIGOT - Professeur des universités - Université de Pau et des Pays de l'Adour - CoDirecteur de these
- Olivier MONDAIN-MONVAL - Professeur des universités - Université de Bordeaux - Examinateur
- Valérie HéQUET - Professeure - IMT Atlantique - Examinateur
- Anne GALARNEAU - Directrice de recherche - Université de Montpellier - Examinateur
The primary objectives of this PhD focused over the syntheses and characterizations of multi-scale inorganic catalysts bearing hierarchical porosity, addressing heterogeneous catalyses. These innovative materials were developed through the integrative chemistry while combining sol-gel process, lyotropic mesophases, and emulsions. The resulting catalysts are self-supporting monoliths, primarily composed of amorphous silica, which acts as a continuous matrix, bearing active nano-sites (such as Co₃O₄, CuO, Cu°, TiO₂, etc.). A significant part of the thesis was dedicated to optimizing the synthetic protocols. Initially, the materials were produced through manual emulsification process, but the research later shifted towards a mechanized formulation. This approach aimed at improving the homogeneity and reproducibility of the formulation step, while fostering industrial high scale production. Additionally, the monolith syntheses required a very slow drying process (1 to 2 months in a desiccator) due to capillary constraints. As a result, part of the research aimed at employing supercritical CO₂ within a commercial apparatus to favor large-scale drying. Given the innovative and complex nature of these ceramics, their characterization at multiple scales was a key focus of the study. A variety of techniques were employed to feature their structure and morphology, including electron microscopy, mercury intrusion, gas sorption, and X-ray diffraction, among others. The catalysts were tested for several reactions, including gas-phase photocatalysis (for the removal of volatile organic compounds) and thermo-activated reactions (such as the oxidation of CO to CO₂).
Calcium-based systems for the anionic polymerization of butadiene: towards a lithium free chemistry
by Thomas LA PORTA (Laboratoire de Chimie des Polymères Organiques)
The defense will take place at 9h00 - IECB - Amphithéâtre Auditorium Institut Européen de Chimie et Biologie (IECB) 2 Rue Robert Escarpit, 33600 Pessac
in front of the jury composed of
- Stéphane CARLOTTI - Professeur des universités - Bordeaux INP - Directeur de these
- PHILIPPE ZINCK - Professeur des universités - UCCS UMR 8181 - UNITE DE CATALYSE ET DE CHIMIE DU SOLIDE, UNIVERSITÉ DE LILLE - Rapporteur
- VINCENT MONTEIL - Directeur de recherche - CP2M, CNRS/Ecole Supérieure de Chimie Physique Electronique de Lyon/Université Claude Bernard - Rapporteur
- NICOLAS ILLY - Maître de conférences - Sorbonne Université - Examinateur
- FRÉDÉRIC PERUCH - Directeur de recherche - CNRS, LCPO - Examinateur
- CHARLOTTE DIRE - Docteure - Manufacture Française des Pneumatiques Michelin - Examinateur
The microstructure and macrostructure of polybutadiene play an essential role in the thermomechanical properties of the material, which is mainly used in tire manufacture. Thanks to its control over the polymerization process and its living character, the anionic polymerization enables a wide variety of polymer architectures to be obtained with precision, thus offering the possibility of tuning the material's properties. However, the synthesis of polybutadiene with a high content of 1,4-trans units remains poorly studied. This thesis proposes calcium-based multi-metallic systems for the controlled and living synthesis of stereospecific 1,4-trans polybutadienes. Lithium-free systems are also proposed as a more sustainable, less toxic and an economically favorable alternative.
ED Entreprise Economie Société
Understanding risk & adoption in individual responses to environmental uncertainty
by Michael MURPHY (BSE - Bordeaux sciences économiques)
The defense will take place at 11h00 - Salle de séminaire de BSE 16 Av. Léon Duguit, 33600 Pessac, France
in front of the jury composed of
- Tanguy BERNARD - Professeur - Université de Bordeaux - Directeur de these
- Isabelle CHORT - Professeur - Université de Pau et des Pays de l'Adour - Rapporteur
- Julie SUBERVIE - Professeur - INRAE CEEM - Rapporteur
- Quentin STOEFFLER - Maître de conférences - Université de Bordeaux, BSE - Examinateur
- David SPIELMAN - Professeur - International Food Policy Research Institute - Examinateur
Individuals must often make decisions under conditions on uncertainty, where the risks associated with their choices are difficult or impossible to quantify. This is particularly true of interactions with the environment where economic agents may lack understanding of complex natural systems. Using three studies of decision-making in the context of food systems in developing economies, this thesis examines individual decisions to change behaviour in response to environmental uncertainty. The chapters of the thesis each explore this topic in a different setting, using a distinct set of analytical techniques. The first chapter analyzes a framed field experiment carried out with farmers in northern Ghana. It focuses on uncertainty surrounding the benefits associated with a set of agroecological practices known collectively as ‘conservation agriculture' techniques. These practices have public good characteristics, but whereas the cost to adoption is immediate, the private benefits to producers accrue over time. The incentivized experiment simulates the farmer's decision to adopt (or not adopt) conservation agriculture practices across seasons. The analysis tests the effectiveness of two approaches to reducing uncertainty around the private benefits associated with adoption, in order to evaluate potential policy approaches to encourage take-up. The second chapter is a secondary panel data analysis looking at the effect of temperature shocks on wheat production in India in the context of the Green Revolution. Uncertainty around weather outcomes has always been a concern of farmers, but understanding how producers adapt- or fail to adapt- to unpredicted weather is an increasingly urgent question in the age of global warming. I apply recent advances in the econometrics of estimating agricultural adaptation to climate to a novel panel dataset and explore how the adoption of Green Revolution technology in the form of high-yielding variety seeds and associated inputs affected farmers' ability to adapt to temperature shocks. The third chapter focuses on consumers, implementing a randomized control trial with low-income urban households in Kenya. The study provides treated households with salient information on a food safety issue- aflatoxin contamination of maize flour- using data from a prior study carried out in the area. The analysis examines how providing participants with information about this hazard affects their subjective risk perceptions, and in turn how this belief updating affects their purchasing behaviour. Across these three different settings, the thesis show how different microeconomic approaches to causal inference can help to understand how individuals respond to a changing environment. These findings aim to improve our understanding of decision-making under uncertainty in contexts where it is rarely studied, and also aim to contribute to a practical evidence base for policymakers in developing countries to design welfare-enhancing policies that enhance citizen access to information and new technologies.
ED Sciences de la Vie et de la Santé
Effect of a plant extracts and omega-3 combination on biological processes involved in age-related cognitive decline
by Marie MARTIN (Nutrition et Neurobiologie Intégrée)
The defense will take place at 15h00 - Amphithéâtre Broca Centre Broca, 146 Rue Léo Saignat, 33000 Bordeaux, France
in front of the jury composed of
- Corinne JOFFRE - Directrice de recherche - Université de Bordeaux - Directeur de these
- Aurélie DE RUS JACQUET - Professeure adjointe - Université Laval - Rapporteur
- Niyazi ACAR - Directeur de recherche - UMR CSGA, INRAE Dijon - Rapporteur
- Virginie AIRES - Maîtresse de conférences - Université de Bourgogne - Examinateur
- Frédéric CALON - Professeur - Université Laval - Examinateur
- Agnès NADJAR - Professeure - Université de Bordeaux - Examinateur
Population's aging represents one of the world's greatest economic, societal and public health challenges. By 2050, more than a billion people worldwide will be aged 60 or over, making it all the more important to age well. In this context, age-related cognitive decline, characteristic of aging, is a critical factor to be considered in order to improve the quality of life of the elderly and prevent loss of autonomy. One promising approach is to propose nutritional strategies. Among existing bioactive nutrients, polyphenols from berries (grapes, blueberries), saffron compounds and omega-3 polyunsaturated fatty acids (n-3 PUFAs) from fish are recognized for their beneficial properties on memory functions. The aim of this thesis was to evaluate the efficacy of a combination of these nutrients in preventing age-related cognitive decline, and to understand the neurobiological mechanisms involved. We targeted the hippocampus, a key brain structure involved in memory processes and particularly impaired during aging. Our main results show that a combination of blueberry, grape, saffron and n-3 PUFA extracts in aged mice effectively prevents both short- and long-term memory impairment. These results are associated with improved neuronal plasticity and neurogenesis, including an increase in DCX, a marker of immature neurons, and up-regulation of signalling pathways involved in synaptic development and function. These beneficial effects can be explained by the modulation of several key biological processes. Indeed, the nutrient combination showed a powerful anti-inflammatory effect by regulating inflammatory lipid mediators, notably pro-resolving oxylipins (18-HEPE, 17-HDHA, 14-HDHA) in adult mice in the setting of acute inflammation, and by inhibiting inflammation-related signaling pathways in aged mice in the setting of chronic inflammation. In addition, it limits oxidation by increasing the activity of antioxidant enzymes such as catalase, superoxide dismutase and glutathione peroxidase in adult mice in response to acute inflammation or treatment inducing accelerated aging, and by down-regulating oxidative stress-related signalling pathways in aged mice. We have also demonstrated that the gut microbiota may play an important role in these effects via the gut-brain axis. In particular, we have demonstrated an increase in bacteria favorable to cognitive health, and a decrease in those associated with inflammation and memory impairment. Finally, this work has enabled us to identify a pool of genes whose expression is increased during aging in the hippocampus and is negatively correlated with cognitive performance. These genes could serve as biomarkers of cognitive decline. Interestingly, their expression is diminished by the combination of nutrients. Thus, our study demonstrates that the combination of berry polyphenols, bioactive saffron compounds and n-3 PUFAs prevents memory impairment by acting on various major neurobiological processes of aging. This multi-targeted nutritional approach has great potential for preventing age-related cognitive decline and promoting healthy aging. These results pave the way for new nutrition-based preventive strategies to combat the effects of cerebral aging
Multi-omics integration to predict growth and protein turnover in developing fruits
by Chloé BEAUMONT (BFP - Biologie du Fruit et Pathologie)
The defense will take place at 14h00 - amphithéâtre Colette & Josy Bové amphithéâtre Colette & Josy Bové Centre INRAE bâtiment B2 71 Avenue Edouard Bourlaux 33140 Villenave d'Ornon
in front of the jury composed of
- Anne GOELZER - Ingénieure de recherche - INRAe - Rapporteur
- Christine DILLMAN - Professeure des universités - Université Paris-Saclay, Faculté des Sciences - Rapporteur
- Marie-Laure MARTIN-MAGNIETTE - Directrice de recherche - INRAE - Examinateur
- Sébastien MONGRAND - Directeur de recherche - CNRS - Examinateur
To unlock barriers for improving plant production, a better understanding of the underlying mechanisms of fruit development is essential, and protein metabolism is at the heart of these processes. Recent work on estimating protein turnover in tomato fruits has highlighted the key role of protein synthesis in establishing the metabolic and physiological traits of development. Additionally, recent laboratory results have shown the importance of proteins in fruit development, as they can effectively predict the relative growth rate (RGR) of fruits among biomass components across a panel of nine species. In this context, this thesis aims to study the stability of proteins in a panel of ten fleshy fruit species selected for their genetic diversity and contrasting characteristics, such as development duration, ranging from 30 to over 200 days, in order to identify generic and unique properties during growth. To achieve this, we have acquired multi-omic data: transcriptomic, proteomic, and metabolomic at ten developmental stages, all expressed in absolute quantification. First, we conducted a comparison of physiological growth data, including fruit development rates and metabolic fluxes estimated through constraint modeling. This work emphasized the importance of nitrogen metabolism, particularly proteins, in the growth-defense trade-off. Next, the obtained transcriptomic data were explored to determine whether transcripts could predict fruit development characteristics such as growth, RGR, and total protein content. Using generalized linear models, the predictions not only proved to be very accurate and reliable but also demonstrated the central role of proteins and cell wall compounds in fruit growth, based on orthologous genes. The following chapter more descriptively explores the proteomic data acquired from the same samples of the nine fruit species panel. This comparative analysis of proteins derived from multi-species orthologous genes reveals key biological functions shared among species, including roles in fruit maturation, regulation of energy metabolism, and processes related to protein synthesis and degradation. Finally, the last section compares protein turnovers calculated from transcriptomic and proteomic data for the nine fruit species with a mathematical model established for tomato fruit. It appears that the synthesis constants are relatively stable across species, while the degradation constants of proteins depend on developmental durations, thus differentiating between herbaceous and tree families. Fruits with slower development, such as trees, possess more stable proteins compared to fast-growing fruits like herbaceous species.
Analysis of the root system architecture responses to reserves availability in grapevine rootstocks using the Archisimple model.
by Mathieu LARREY (Ecophysiologie et Génomique Fonctionnelle de la Vigne)
The defense will take place at 9h00 - Amphithéâtre ISVV 210 Chemin de Leysotte, 33140 Villenave-d'Ornon
in front of the jury composed of
- Philippe VIVIN - Chargé de recherche - INRAE Bordeaux - Directeur de these
- Valerie LEGUE - Professeure - Université Clermont Auvergne - Rapporteur
- Guillaume LOBET - Professeur - Université Catholique de Louvain - Rapporteur
- Marie-Béatrice BOGEAT TRIBOULOT - Chargée de recherche - INRAE Nancy - Examinateur
- Christophe NGUYEN - Directeur de recherche - INRAE Bordeaux - Examinateur
- Frédéric REES - Chargé de recherche - INRAE Versailles-Grignon - Examinateur
Improving the recovery rate of young grafted grapevines in order to limit their future decline in vineyards is a major challenge for nurseries and grape growers. This requires a better understanding of their rooting capacities and the establisment of their root system architecture (RSA) after implantation, in particular in relation to their efficiencies in soil prospection and soil resources capture. Many genetic, physiological or environmental factors modify these processes in grapes. Among them, the nutritional status of the mother plant material could be decisive. The aim of this thesis is to characterize, in young vines mostly under controlled conditions, the influence of the initial availability of carbon and nitrogen reserves on rhizogenesis, growth and plasticity of the morphological traits of the root system for different Vitis rootstock genotypes contrasted in terms of root development. In particular, it will be investigated whether the processes of root emission, elongation and branching are differently affected by the local carbohydrates content and the nature of the graft. These data will also be used to calibrate a first model of root system development for the vine. Such functional-structural root model is a promising way to overcome in situ root phenotyping issues and can serve as a basis for the development of root ideotypes in silico by highlighting the underlying genetic mechanisms and parameters that are most likely influence RSA.
ED Sciences Physiques et de l'Ingénieur
Contribution to the study of factors influencing risk perception by aircraft maintenance mechanics
by Raphael CHIRAC (Laboratoire de l'Intégration du Matériau au Système)
The defense will take place at 14h00 - amphitéatre ENSC 351 cours de la Libération CS10004 33405 , TALENCE CEDEX
in front of the jury composed of
- Jean-Marc ANDRE - Professeur des universités - Université de Bordeaux - Directeur de these
- Benoit EYNARD - Professeur des universités - Université de technologie de Compiègne - Rapporteur
- André TRICOT - Professeur des universités - Université Paul Valéry - Rapporteur
- Franck CAZAURANG - Professeur des universités - Université de Bordeaux - CoDirecteur de these
- Herimanana ZAFIHARIMALALA - Docteure - Human design Group - Examinateur
- Arturo MARTINEZ-GRACIDA - Docteur - Bombardier - Examinateur
This study of risk perception by aircraft maintenance mechanics is based on the observation that the safety of flight operations is mainly considered through a regulatory framework in order to ensure a high level of safety. However, this does not allow us to understand how mechanics manage risks on a day-to-day basis. Thus, this work was carried out to answer the following 3 research questions: How do aircraft mechanics perceive risks in terms of the 3 risk characteristics (probability, severity, immediacy)? What factors might explain the variance in risk perception? Is there a difference between the perception of risks impacting mechanics' health and the perception of risks impacting aircraft airworthiness? In order to identify the hazards to which mechanics are exposed during their activity, we established a taxonomy. An experimental approach was followed to assess its relevance and completeness and to validate it. The results show that the 10 engineers and 20 mechanics of our panel share a common representation of the hazard. We also identified confusion between the notions of hazard, risk and safety measure. We conclude that it is important for the operators to share a clear definition of these notions. From our point of view, this corresponds to the main basis for ensuring safe maintenance operations and the airworthiness of the aircraft. Then, we designed a model of risk perception adapted to the aircraft maintenance domain, which was tested on a new panel of 27 aeronautical mechanics. We concluded that one of the major challenges is to ensure that mechanics share a common perception of risk that reflects the reality of the activity. To achieve this, we highlighted the impact of three sets of factors on risk perception: the intrinsic characteristics of a hazard; the mechanic's knowledge and experience; and the operational context. For this purpose, effective communication and training in risk management are needed, so that mechanics obtain an applicable risk perception to the operational context.
Study of a Radio Frequency Front End (RFFE) based on Walsh sequences for the 5G standard and beyond.
by Pierre FERRER (Laboratoire de l'Intégration du Matériau au Système)
The defense will take place at 9h00 - Amphithéâtre Jean-Paul DOM 351 Cours de la Libération, 33405 Talence Cedex, France
in front of the jury composed of
- François RIVET - Maître de conférences - Université de Bordeaux - Directeur de these
- Ramesh HARJANI - Full professor - University of Minnesota Twin Cities - Rapporteur
- Antoine FRAPPE - Associate Professor - Junia - Rapporteur
- Daniela DRAGOMIRESCU - Professeur des universités - LAAS - Examinateur
- Doug SMITH - Ingénieur - Austin, Texas - Examinateur
- Nathalie DELTIMPLE - Professeur des universités - Université de Bordeaux - Examinateur
- Yann DEVAL - Professeur des universités - Université de Bordeaux - CoDirecteur de these
- Hervé LAPUYADE - Maître de conférences - Université de Bordeaux - Co-encadrant de these
Mobile telecommunications services require a continuous and exponential increase of their capacities. To meet this demand, transceivers must be highly digital and flexible. A promising solution to achieve this goal is to use Walsh sequences to generate arbitrary waveforms. This is based on the principle of a "software defined radio" transmitter comprising a high-speed, high-resolution DAC. Thus, the studied architecture consists of a very wide band DAC (0 to 6 GHz) very efficient for a direct RF amplification with 5G phase 1 applications as target and a Hilbert Transform based signal shaping for a millimetric frequency upconversion (beyond 20GHz) without resorting to complex filtering techniques with 5G phase 2 or 6G applications as target. The objective of this work is to propose a new transmitter topology that responds to issues such as: massive data conversion (very high throughput), carrier aggregation and dynamic linearization of an RF front end. The candidate will demonstrate the proposed architecture using ST Microelectronics 28nm FDSOI technology. The work will include high level simulations (MatLab), architecture sizing, schematic simulations, mask designs, circuit fabrication, design of a suitable measurement environment and comprehensive measurements.
ED Sciences et environnements
Soil biodiversity and functioning in vineyards
by Pierre BLONDEL (SAVE - Santé et Agroécologie du VignoblE)
The defense will take place at 13h30 - Amphithéâtre 210 Chemin de Leysotte 33140 Villenave-d'Ornon
in front of the jury composed of
- Gwenaëlle LASHERMES - Chargée de recherche - INRAE, UMR FARE - Rapporteur
- Virginie BALDY - Professeure des universités - Institut Méditerranéen de Biodiversité et d'Ecologie Marine et Continentale (IMBE) - Examinateur
- Jérôme CORTET - Professeur des universités - Université Paul-Valéry Montpellier 3 - Rapporteur
- Benjamin PEY - Maître de conférences - INP-ENSAT - Examinateur
Winegrowing practices are impacting biodiversity, in relation with the perennial nature of the crop and the simplification of the landscape. Soils are very often weeded or tilled, and vineyards are heavily dependent on phytosanitary treatments. However, these practices have an impact on biodiversity and the ecosystem services this biodiversity provides. One of the major objectives of the agroecological transition of vineyards is to find agroecological and integrated management practices based on the preservation of biodiversity and the services it provides. Thus, organic farming has developed, but how effective is it? Are there efficient practices that improve the functioning of winegrowing soils? What about the effects of landscape diversity? Our aim was to assess the impact of viticultural practices, at the field and inter-row levels; but also to evaluate the influence of the surrounding landscape. We focused on the activity of soil microorganisms and the diversity of mesofauna, the first links in trophic networks; these communities are at the heart of processes essential to soil functionality. To this end, our work was organized around the Bacchus network, a network of pairs of differently managed vineyards. The network is located in a relatively homogeneous wine-growing landscape, with a gradient of semi-natural habitats and organic farming areas. Our studies focused on the effect of practices on soil microorganisms (chapter 1), with analysis of extracellular enzyme activities and respiration of carbon substrates. We observed that surface soil organic matter levels increased in organically managed plots, boosting the enzymatic activities of microbial communities. On an intra-plot scale, inter-row grassing resulted in more complex carbonaceous substrate-breathing communities. These two management parameters also favored nutrient recycling and nitrogen availability. Organic management also benefits springtail communities (Chapter 2), by increasing organic matter levels. The intensity of pesticide use, but above all the number of pesticides applied, reduces the diversity of Collembola communities, whereas soil copper contamination showed no impact on our studied communities. On an intra-plot scale, tillage also favored the diversity of these communities. Apart from practices, pH and seasonality greatly modulated community diversity, with an unfavorable summer period. Landscape diversity surrounding vineyards (Chapter 3) also appears to play a role in soil biodiversity. The partitioning of β-diversity into its two components (turnover and nestedness) showed contrasted hemiedaphic springtail communities between organic and conventional plots. Increasing the area of organically-farmed vineyards in the landscape enhanced diversity in organic plots, but not in conventional plots, which were favored by the proportion of surrounding semi-natural habitats. These studies highlight the importance of soil management in viticulture, avoiding acidification and, above all, the loss of organic matter. Organic management of vineyards preserves biodiversity to a greater extent, by reducing the intensity of pesticide use and, in this case, the effects of copper contamination are fairly slight, while the results for grass management are more contrasted. Finally, the landscape diversity is an important element to consider in the management of viticultural practices and its effect on soil biodiversity.
ED Sociétés, Politique, Santé Publique
GENETIC DETERMINANTS OF MRI-MARKERS OF VASCULAR BRAIN AGING
by Yasaman SABA (Bordeaux Population Health Research Center)
The defense will take place at 14h00 - Salle de conférence 146 rue Léo Saignat,, 33076, Bordeaux
in front of the jury composed of
- Elisabeth TOURNIER-LASSERVE - Professeure des universités - praticienne hospitalière - Université Paris Cité - Examinateur
- Hugues CHABRIAT - Professeur des universités - praticien hospitalier - Université Paris Cité - Rapporteur
- Cecilia SAMIERI - Directrice de recherche - Université de Bordeaux - Examinateur
- Stephanie DEBETTE - Professeure des universités - praticienne hospitalière - Université de Bordeaux - Directeur de these
- Helena SCHMIDT - Professeur - Université de Graz - CoDirecteur de these
- Israel FERNANDEZ-CADENAS - Docteur - IIB Sant Pau and Sant Pau Hospital - Rapporteur
Over the last century, life expectancy has increased dramatically, contributing to a sharp increase in the number of patients with common neurological disease, especially stroke and dementia. Mounting evidence suggests that early life factors, including genetic factors, play a crucial role in the occurrence of such diseases. Cerebral small vessel disease (cSVD) is a major cause of stroke, cognitive decline and dementia. cSVD is most often covert, detectable on brain images in the absence of clinical manifestations. Brain magnetic resonance imaging (MRI) markers of cSVD, which can be measured non-invasively in large population, can provide crucial insights into the cause of late-life neurological diseases. White matter hyperintensities (WMH), lacunes, cerebral microbleeds, and perivascular spaces are the most commonly studied MRI-markers of cSVD, while diffusion tensor imaging (DTI) offers new opportunity to explore susceptibility to cSVD across the lifespan. Deciphering these genetic risk factors of cSVD, including in early life, is a powerful tool to decipher molecular mechanisms leading to this disease. In this thesis, we explored the genetic determinants of MRI-markers of cSVD in the general population across the lifespan, by conducting large collaborative meta-analyses of genome-wide association studies (GWAS) in up to 58,403 participants from the general population. First, we conducted a GWAS of WMH stratified on hypertension status. Our results shed new light into modifying effects of high blood pressure on genetic susceptibility to WMH. Second, we examined the genetic underpinnings of an emerging DTI marker, peak width of skeletonized mean diffusivity (PSMD), by conducting a first GWAS of PSMD, across the lifespan. We identified up to 25 novel genetic risk loci for PSMD, with good effect size correlation across European and East-Asian ancestries. Additionally, in a whole-exome association study (derived from whole exome sequencing), rare variants and burden of rare loss-of-function or singleton variants in 4 different genes were associated with PSMD. Genetically determined larger volume of WMH was associated with higher PSMD from early childhood to older age. Moreover, common PSMD risk loci were enriched in genes expressed in fetal brain endothelial cells. In conclusion, this work provides new insights into complex genomics of cSVD across the lifespan, across ancestries, and in interaction with hypertension, the most common risk factor of cSVD. These results are informative for the development of efficient preventive and therapeutic strategies for cSVD and its complications, a major public health challenge.