ED Mathématiques et Informatique
I/O characterization of HPC applications for scheduling
by Alexis BANDET (LaBRI - Laboratoire Bordelais de Recherche en Informatique)
The defense will take place at 9h00 - Ada Lovelace 200 Av. de la Vieille Tour, 33405 Talence
in front of the jury composed of
- Guillaume AUPY - Chargé de recherche - Centre Inria de l'Université de Rennes - Directeur de these
- Francieli ZANON-BOITO - Maîtresse de conférences - Universite de Bordeaux - CoDirecteur de these
- Samuel THIBAULT - Professeure des universités - INRIA, University of Bordeaux - Examinateur
- Shadi IBRAHIM - Chargé de recherche - Centre Inria de l'Université de Rennes - Examinateur
- François TRAHAY - Professeur des universités - Télécom SudParis - Rapporteur
- Soraya ZERTAL - Professeure des universités - ISTY - Institut des sciences et techniques des Yvelines - Rapporteur
High-performance computing (HPC) is an integral part of today's scientific workflow. Aircraft conception, weather simulation, medicine, or astronomy research are only a few of its many fields of operation. Recent convergence between HPC and Big Data involve an important stress on Input/Output (I/O) infrastructures. Unlike compute resources, which are granted with exclusive access to an application, I/O resources are typically shared by all applications, which may hence cause interference in each others' I/O performance. The non-deterministic nature of this slow-down is a problem for resource management. Moreover, slowing down I/O often means the expensive, power-hungry compute resources are left idle while applications wait for data operations. To mitigate this issue, scheduling of applications on I/O resources should be applied. In this thesis, we focus on profiling and characterize application's I/O to provide the scheduler the information it needs to make good choices. We aim to identify the minimal information we have to know about an application. Existing tools for I/O monitoring, like Darshan and Tau, were not designed to handle these new challenges. They provide way more information than needed, creating an extensive, but in our case useless, amount of data. In this thesis, we firstly worked to identify the minimal required knowledge about applications to schedule distributed I/O resources, such as OSTs and I/O nodes. For that, we designed simple, but effective, scheduling heuristics that require different levels of application knowledge. We demonstrated that even basic information such as the number of resources that maximize the I/O throughput is enough to observe improvements through I/O scheduling. We then focused on proposing a way to obtain this information. We used some machine learning techniques to find the stripe count that maximizes the I/O bandwidth of an application. As model input, we used I/O applications parameters such as request size, number of processes performing I/O operations, and access pattern. Furthermore, we used feature pruning to better understand the role of each parameter in our prediction. We conclude that all features remain important for good prediction, but with different orders of magnitude. Lastly, we worked on profiling the temporal behavior of HPC applications, which are known to behave in an alternation of compute and I/O phases. This periodicity is an important component because it creates a temporal but huge stress on I/O systems, called a burst. Related work has showed that information about the frequency of I/O phases can be used for efficient scheduling of accesses to the shared file system. For that purpose, we developed FTIO, a tool that takes advantage of Fast Fourier Transform algorithm to predict application periodicity. In addition to the frequency of I/O phases, we proposed metrics to verify if the application is periodic, which allow for verifying if the obtained frequency is correct. Taking into account the temporal behavior is an important part of this thesis' contributions, as it is common for applications to change behaviors during their execution.
ED Sciences Chimiques
Theoretical study of the vibrational relaxation of water by collisions with some of the most abundant collision partners present in evolved stars' atmospheres
by Ricardo Manuel GARCIA VAZQUEZ (Institut des Sciences Moléculaires)
The defense will take place at 10h00 - Amphis C 351 Cours de la Libération, Bâtiment A29, 33405 TALENCE cedex
in front of the jury composed of
- Thierry STOECKLIN - Directeur de recherche - Institut des Sciences Moléculaires, Université de Bordeaux - Directeur de these
- Majdi HOCHLAF - Professeur des universités - Université Gustave Eiffel - Rapporteur
- Ugo ANCARANI - Professeur des universités - Université de Lorraine - Rapporteur
- Astrid BERGEAT - Maîtresse de conférences - Institut des Sciences Moléculaires, Université de Bordeaux - Examinateur
- Otoniel DENIS-ALPIZAR - Assistant professor - Facultad de Ciencias, Universidad de Chile - Examinateur
- Alexandre FAURE - Directeur de recherche - Institut de Planétologie et d'Astrophysique de Grenoble, Université de Grenoble Alpes - Examinateur
- Philippe HALVICK - Chargé de recherche - Institut des Sciences Moléculaires, Université de Bordeaux - Examinateur
- Fabrice HERPIN - Astronome - Laboratoire d'Astrophysique de Bordeaux, Universite de Bordeaux - Examinateur
The existence of water molecules in space was confirmed in 1964 when the presence of spectral lines of water vapour was detected in the atmosphere of O-rich stars. Subsequently, water has been identified as a prevalent molecule across the universe. Rotational lines have now been observed in a variety of environments, including molecular clouds and star-forming regions. For many years, the astrophysical models used to model the physical conditions prevailing in these environments have only considered the water molecule rotation. However, studies of warmer environments, such as the envelopes and atmospheres of evolved stars, have led to the detection of vibrationally excited H2O molecules. Also, vibrational de-excitation rate coefficients for the collisions of H2O with the most abundant elements in the interstellar media, including H2, He, H and e- are needed to model these environments. Unfortunately, most of the currently available theoretical collisional data are limited to rotational transitions. Experimental studies suffer from the same limitation but offer the opportunity to validate the accuracy of the theoretical model employed. Furthermore, when the relevant astrophysical systems are too challenging to measure, experimentalists often turn to similar systems as a proxy for the systems of interest. The present work, in addition to providing new vibrational data, then also includes rotational (de)excitation studies of systems which are compared with available experimental data. This is the case of the rotational (de)excitation of H2O, HDO and D2O molecules by collisions with Ne. An excellent agreement between theory and a new recent experiment is obtained for the collisions of HDO with Ne validating the new PES developed for this system. which was then used to infer collisional propensity rules for these collisions. Similarly, as the differential cross section (DCS) offers the most detailed information about a collisional process, the collisions of H2O with Ne, Ar and Xe are also studied, and a very good agreement is obtained with the experimental results. The second main objective of this work is to study the collisional relaxation of vibrationally excited levels of water. The first systems to be studied using the Rigid Bender Close Coupling method (RBCC) were the bending relaxation of H2O by collision with Ar and of D2O by collision with He and Ar atoms. The subsequent comparison with experimental data allowed validating this approach. The last part of this manuscript is dedicated to the vibrational relaxation of H2O by collisions with the most abundant astrophysical collider, H2. The RBCC method is extended to include all the vibrational modes of H2O and the collisions with a linear molecule instead of an atom. A new code, called "Divitas," is developed allowing to perform both close coupling (CC) and coupled states (CS) calculations to study the rovibrational relaxation of triatomic molecules, such as H2O, in collision with linear molecules, such as H2. The new code is first employed to investigate the bending relaxation of H2O by collisions with H2 utilising the CC method and a new set of rate coefficients including ro-vibrational transitions is produced. Our global bending relaxation rate is in excellent agreement with the experimental value reported at 295 K. Nevertheless, the high computational cost of the CC method represents a significant obstacle to the study of more excited vibrational states. To circumvent this difficulty, the use of the CS approximation is first validated by comparison with the CC results. It is then subsequently employed to investigate the relaxation of the first bending overtone utilising a H2O basis set incorporating the first five vibrational states of the water molecule.
Nanoscale chemical, structural and morphological imaging in aqueous medium: Development of optical nanosources by electrodeposition and characterization of amyloid aggregates
by Yuhan HUANG (Institut des Sciences Moléculaires)
The defense will take place at 10h00 - Salle de Conférence (3ème EST) Bâtiment A12 — 351 Cours de la Libération 33405 TALENCE
in front of the jury composed of
- SEBASTIEN BONHOMMEAU - Maître de conférences - Université de Bordeaux - Directeur de these
- Emmanuel MAISONHAUTE - Professeur des universités - Sorbonne Université - Rapporteur
- Katrin F. DOMKE - Full professor - Max Planck Institute for Polymer Research - Rapporteur
- Bernard HUMBERT - Professeur des universités - Nantes Université - Examinateur
- Christine GRAUBY-HEYWANG - Professeure des universités - Université de Bordeaux - Examinateur
Tip-enhanced Raman spectroscopy (TERS) is a powerful technique for characterizing the structure and composition of amyloid fibrils, which is fundamental for a deeper understanding of associated diseases. This thesis dissertation focuses on two major scientific axes in TERS, namely the fabrication of efficient TERS probes and their use for nanoscale chemical and structural imaging in ambient air and in aqueous medium. After general theoretical and technical information about TERS, it introduces the development of gold-coated atomic force microscopy (AFM) tips using potentiostatic electrodeposition for applications in TERS. This cost-effective method allowed robust TERS probes to be fabricated, and enables TERS mapping of a graphene oxide flake and a carbon nanotube with lateral spatial resolution below 20 nm. Furthermore, an original method based on bipolar electrodeposition to prepare gold-coated AFM tips used as TERS probes is also described. This wireless method is simple to implement, cost-effective, and allows for the parallel fabrication of several TERS tips with good repeatability. The TERS activity was confirmed by imaging graphene oxide flakes with high spatial resolution (below 10 nm). TERS tips fabricated by bipolar electrodeposition method were then selected to perform total-internal-reflection TERS (TIR-TERS) imaging of amyloid-β (Aβ(1–42)-L34T) fibrils in water. TIR-TERS imaging reveals the predominant parallel β-sheet secondary structure of Aβ(1–42)-L34T fibrils and the nanoscale spatial distribution of tyrosine, histidine, and phenylalanine aromatic amino acids. Conclusions drawn from the TERS experiments in water corroborate and enrich previous studies, confirming that hydration has only a marginal impact on the structure of such amyloid fibrils. Therefore, this first TIR-TERS study in liquid environment opens fascinating perspectives for future applications in biology.
Development of an electrochemical method for the redox state of quinones measurement at different levels of cellular integration in the yeast Saccharomyces cerevisiae
by Mélanie MARTINS PINTO (Institut de Chimie & de Biologie des Membranes & des Nano-objets)
The defense will take place at 9h30 - Salle de conférence IBGC IBGC Institut de Biochimie et Génétique Cellulaire UMR 5095 1 rue Camille Saint Saëns 33077 Bordeaux Cedex
in front of the jury composed of
- Stéphane ARBAULT - Directeur de recherche - Université de Bordeaux - Directeur de these
- Anne DEVIN - Directrice de recherche - IBGC UMR 5095 - CoDirecteur de these
- Petra HELLWIG - Professeure des universités - Université de Strasbourg - Rapporteur
- Fabien PIERREL - Directeur de recherche - Laboratoire TIMC UMR 5525 - Rapporteur
- Bertrand BEAUVOIT - Maître de conférences - INRAE UMR 1332 - Examinateur
- Frédéric LEMAITRE - Professeur des universités - Sorbonne Université - Examinateur
Cellular energy metabolism includes all the metabolic pathways involved in the conversion of the chemical bond energy stored in carbon sources available in the environment into energy in the form of adenosine triphosphate (ATP), which is essential for carrying out key cellular functions. It is also involved in maintaining a redox potential compatible with the cell's enzymatic reactions, through the reoxidation of reduced coenzymes, such as NADH. In aerobic cells, mitochondria couple the reoxidation of reduced molecules with the synthesis of the majority of cellular ATP through the oxidative phosphorylation system (OXPHOS). The cellular redox potential (defined by the NADH/NAD+ ratio) is thus an important parameter for cellular metabolism, as it regulates redox reactions associated with ATP production as well as cellular homeostasis in response to oxidative stress. However, determining the cellular redox potential is challenging due to its dual compartmentalization (cytosolic and mitochondrial) and the fact that most cellular NADH is bound to enzymes. Quinones are lipophilic molecules located in the inner mitochondrial membrane (Q6 in the yeast Saccharomyces cerevisiae) that initiate electron transfer at the beginning of the respiratory chain of OXPHOS, from dehydrogenases to complex III. Their redox state depends on the respiratory chain's capacity to reoxidize available substrates in the cell and, more generally, on the function of OXPHOS, which lies at the core of cellular energy metabolism. The objective of this work was to develop an electrochemical technique to monitor in real-time the reduction rate of quinones using a redox mediator, coenzyme Q2, in conjunction with mitochondrial respiration measurements. We developed this technique in the yeast S. cerevisiae at three levels of integration: isolated mitochondria, permeabilized spheroplasts, and intact spheroplasts. Our results with isolated mitochondria clearly demonstrate that the redox state of quinones is a marker of mitochondrial metabolism, dependent on the respiratory substrate, the mitochondrial bioenergetic state, and the function of the Krebs cycle. Studies on permeabilized spheroplasts revealed different regulations of quinone reduction rates compared to those observed at the first level of integration, showing an influence of a preserved cytosolic environment on the measured parameter. Preliminary studies on intact spheroplasts suggest a connection between cellular metabolism and the quinones redox state. Overall, this work provides additional insights into the understanding of the flux/force relationships regulating OXPHOS activity and the influence of the cellular environment on mitochondrial metabolism.
ED Entreprise Economie Société
The influence of building institutional capacity in the quest of economic development in subsaharan Africa countries
by Moulaye OWONA BINDZI (BSE - Bordeaux sciences économiques)
The defense will take place at 14h00 - salle des thèses université 16 avenue Léon Duguit 33600 Pessac
in front of the jury composed of
- Désiré AVOM - Professeur des universités - UNIVERSITE DE YAOUNDE 2 - Examinateur
- François COMBARNOUS - Professeur des universités - Université de Bordeaux - Examinateur
- Sophie BRANA - Professeure des universités - Université de Bordeaux - Examinateur
- ITCHOKO MOTANDE MONDJELI MWA NDJOKOU - Professeur des universités - UNIVERSITE DE MAROUA - Rapporteur
- Françoise OKAH EFOGO - Maître de conférences - UNIVERSITE D'EBOLOWA - Examinateur
- Roger Antoine Pepin TSAFACK NANFOSSO - Recteur d'université - UNIVERSITE DE DSCHANG - Rapporteur
This thesis aims to address the influence of institutional quality on economic development regarding econometric evidence of the link between Cameroon long-term growth's bottlenecks seen as a proxy and its cause roots. Sometimes related to good governance or to the rules that lead the process of managing public organizations efficiently, institutional quality and pub-lic governance seems to be wide and multi-variable concepts that could be broken down as follows: voice and accountability, government effective-ness, regulatory quality, rule of law, economic freedom, civil liberties, in-come inequality, corruption. Conversely to countries with a great political and cultural fabric, building state capacity is generally recognized as one of the key drivers that might explain the poor political, economic, environmental and social assessments of sub-Saharan countries' public policies. The study will generally assess the opportunity and the relevance of public institutions on sub-Saharan countries to foster economic development and highlights for African countries governance performance and shortcomings as well as the fitting and consistency of public institutions regulation with national and regional development strategy or frameworks. Therefore, the perpetuation of those missing links that prevent planning and control and implementation tools to be efficient, in a bureaucratic setting, can easily be related to the average performance of public institutions. As a matter of fact the situation that prevails for public institutions pave the way for inefficient or ineffective private institutions either in structured and organized the market or in the dominant African informal sector. Along the way of the implementation of the African Union 2063 development agenda, the strategic and operational bottlenecks encountered by all these stakeholders oblige African governments to reconcile the human, financial and time-bound ressources available with the smart objectives dedicated to institutional quality. Given this background, the added value of this research product is to establish the scientific relationship that may exist between high institutional quality standards and the current and prospective economic development of subsaharian African countries. The findings based on econometric models show the clear view and heterogeneity of those claimed by the defenders of the institutional economic dis-course. Furthermore they go deeper on the understanding of the strengths and weaknesses of the variables chosen to weight both institutional quality, public governance and economic development of a country. Political economy solutions and ways forward are given to light up the path through the regional or nationwide existing frameworks.
ED Sciences de la Vie et de la Santé
Conformational dynamics and interactions of eIF4B IDR and its phosphomimetic mutants
by Pascale SARKIS (Acides nucléiques : Régulations Naturelles et Artificielles)
The defense will take place at 14h00 - Auditorium 2, Rue Robert Escarpit - 33607 PESSAC - France
in front of the jury composed of
- Mikayel AZNAURYAN - Chargé de recherche - Université de Bordeaux - Directeur de these
- Isabel ALVES - Directrice de recherche - CNRS, université de bordeaux - Examinateur
- Sigrid MILLES - Professeure - FMP - Rapporteur
- Andrea SORANNO - Professeur - WUSTL - Rapporteur
The structure-function paradigm defines that protein function is determined by its structure, and detailed structural knowledge provides critical insights into its functional mechanisms. This paradigm has been challenged with the intrinsically disordered proteins (IDPs) that lack a stable structure, yet they are functional under physiological conditions. Eukaryotic translation initiation factor 4B (eIF4B) is an IDP involved in the regulation of translation initiation in eukaryotes. As an essential co-factor of RNA helicase eIF4A, eIF4B is particularly important for translation of mRNAs with long and structured 5' untranslated regions. It contains several defined functional domains/regions, including the structured N-terminal RNA recognition motif (RRM) domain, the disordered DRYG region, enriched with aspartate, arginine, tyrosine and glycine and the disordered C-terminal arginine-rich motif (ARM) region. While the RRM and ARM domains mediate RNA binding, the DRYG region is essential for eIF4B self-association. eIF4B is overexpressed in cancer cells, and may influence stress granule formation. The cellular activity of eIF4B is regulated by phosphorylation, notably at Ser406 and Ser422 residues. Despite its importance, only the well-structured RRM domain has been characterized at atomic level. The molecular details of its large intrinsically disordered region (IDR) are still unknown, as proteins of this nature are difficult to characterize due to their conformational heterogeneity and dynamic behavior. During my PhD work I had four objectives: i) structural characterization of eIF4B IDR in its monomeric state; ii) characterization of conformational dynamics and interactions of eIF4B IDR on the molecular level upon oligomerization and upon condensation on the mesoscopic scale; iii) investigation of conformational dynamics of eIF4B IDR upon RNA interactions; iv) analysis of the impact of key phosphomimetic mutations on eIF4B protein - protein interactions, eIF4B condensation and eIF4B - RNA interactions. Using single-molecule Förster resonance energy Transfer spectroscopy (smFRET) I studied the eIF4B IDR as a monomer, demonstrating its non-uniform conformational behavior and flexibility, with different regions showing varying degrees of compactness and dynamics. Although the DRYG region is disordered, it is surprisingly compact, whereas the CTR is more expanded and flexible. These characteristics are largely dictated by the specific sequence composition of each subregion. smFRET also enabled probing eIF4B oligomerization behavior and associated protein conformational changes. Increasing the protein concentration above certain thresholds lead to eIF4B phase separation, which was studied by dedicated phase separation assays. Altogether, these experiments enabled mapping of the self-association landscape of eIF4B, which represents a complex transition from monomers to oligomers to condensed droplets. Interestingly, phosphomimetic mutations, such as S406E and S422E minimally affect eIF4B oligomerization, but considerably reduce the phase separation propensity. Finally, I used a combination of smFRET and NMR experiments to investigate eIF4B - RNA interactions, confirming that binding primarily involves the 332 to 457 region (overlapping with previously identified ARM region), which undergoes compaction upon RNA binding. The binding is ionic strength dependent, suggesting that electrostatic interactions are the main driving force, while sequence specificity towards guanosine-containing RNAs, indicates additional π-π stacking interactions. Importantly, the Ser406 and Ser422 phosphomimetic mutations within the RNA binding region significantly affect eIF4B-RNA binding affinity. Altogether, this work provides a detailed molecular understanding of conformational behavior of eIF4B and mechanisms of its interactions.
Correction of physiological movements without external device in PET imaging: applications to low statistical acquisitions for hepatic radioembolization and cardiology.
by Alizée SERIEYSSOL (Institut de neurosciences cognitives et intégratives d'Aquitaine)
The defense will take place at 14h00 - Salle de conférence 146, rue Léo Saignat 33076 Bordeaux cedex Site Carreire de l'Université de Bordeaux Batiment CARF
in front of the jury composed of
- Frédéric LAMARE - Docteur - Université de Bordeaux - Directeur de these
- Catherine CHEZE-LE-REST - Professeure des universités - praticienne hospitalière - Centre Hospitalier Universitaire de Poitiers - Rapporteur
- Thomas CARLIER - Docteur - CHU de Nantes -Hôtel-Dieu - Rapporteur
- Laurence BORDENAVE - Professeure émérite - Université Bordeaux - Examinateur
Positron emission tomography (PET) is an essential imaging modality for diagnosis and therapeutic follow-up in oncology. Physiological motion can degrade image quality and affect the diagnostic accuracy and quantification of PET images. This research program focuses on the detection of physiological motion (respiration and cardiac beating) without the use of an external device for very specific clinical applications. Methods to compensate for these movements will be developed to reconstruct an image corrected for these effects. Two clinical applications have been identified to evaluate the implemented methods. The first concerns hepatic radioembolization based on 90Y PET imaging, which requires the development of methods to detect and correct for respiratory motion for data with very low counting statistics. The second is 18F-FDG cardiac PET imaging, involving the development of a method for the dual detection of respiratory and cardiac movements, as well as methods for compensating for these two physiological movements. The results obtained with the proposed detection methods are compared with those obtained with external devices: a bellow (46-265679G-1, GE HealthCare) for the respiratory signals and an electrocardiogram (ECG) for the cardiac signal. Two correction methods are proposed for hepatic radioembolization and their impact on post-treatment dosimetry was evaluated in comparison with results obtained without the use of correction methods. The first method developed consists in keeping only the quiescent phase of the respiratory cycle, while the second uses all the statistics, proposing a rigid registration between all the respiration phases. Two other methods have been implemented for cardiology, based on the estimation of 3D deformation vectors obtained from cardiac and respiratory triggers calculated with the proposed detection method. The first method estimates these deformation vectors through a rigid registration between the images of each respiratory cycle, while the second method uses the different volumes of the heart. In this method, 3D deformation vectors are calculated by identifying the end diastolic and end systolic volumes. The efficacity of these methods is evaluated by comparing the images obtained using these methods with the non-motion-corrected images, as well as with the image reconstructed with the correction method used in clinical routine on PET/CT cameras (Q.Static algorithm, General Electric HealthCare). The obtained results demonstrate a real improvement in terms of image quality, with better results for cardiological images than those obtained with the correction method used in clinical routine. Dosimetric results obtained with both correction methods for Yttrium-90 data show an increase of the tumor dose.
Analysis of SMN-1 protein role in neuronal development/survival by cell-specific transcriptomic and proteomic analyses in C. elegans
by Sara SAVAHELI (Acides nucléiques : Régulations Naturelles et Artificielles)
The defense will take place at 14h30 - auditorium of Bâtiment bordeaux biologie santé 2, rue Dr Hoffman-Martinot 33000 Bordeaux
in front of the jury composed of
- Denis DUPUY - Directeur de recherche - Université de Bordeaux - Directeur de these
- karine MASSE - Maître de conférences - Université de Bordeaux - Examinateur
- mirko FRANCESCONI - Directeur de recherche - Laboratoire de Biologie et Modélisation de la Cellule ENS de Lyon - Rapporteur
- Nuria FLAMES BONILLA - Directrice de recherche - Molecular and Cellular Pathology and Therapy, Developmental Neurobiology - Rapporteur
- Cindy DEGERNY - Professeur des universités - université Paris-Saclay - Examinateur
Spinal muscular atrophy (SMA) is a neuromuscular disease mainly caused by mutations in the SMN1 gene (Survival of Motor Neuron 1). SMN1 is ubiquitously expressed and controls the assembly of small nuclear ribonucleoproteins (snRNPs), essential for pre-mRNA splicing. What makes some MNs more sensitive to SMN1 depletion is still not fully understood. We are exploring the molecular origins of the distinct sensitivity of different neurons to loss of SMN1 in a C. elegans model using three approaches: -We used neuron-specific RNAi to selectively knock-down the SMN1 homolog, smn-1, in motor neurons or in touch receptor neurons (TRNs) in strains that express fluorescent reporters in the desired cells. In both neuron types, we observed that when smn-1 is silenced, neurons that are born post embryonically appear less robustly than in controls, while neurons which are born embryonically display axonal degeneration that precedes neuronal disappearance. The results suggest the role of SMN both in neural development, and maintenance. -We expressed SMN-1 fused to TurboID in MNs and TRNs to identify potential differential protein interactions involved in neuron development/survival We captured some of the known interactions of SMN-1 in both neuron types as well as other proteins knowns to participate in RNA processing. Analysis of the mass spectrometry results are still ongoing. -Targeted TRN and VD/DD neurons have been isolated to generate cell-specific cDNA libraries for transcriptome sequencing using Nanopore. We are in the process of preparing sufficient cDNAs from isolated cells to reach the needed input for Nanopore sequencing, and identifying downstream transcriptomic changes following SMN-1 depletion.
Xenopus vestibular network adaptation throught metamorphosis
by Gabriel BARRIOS (Institut de neurosciences cognitives et intégratives d'Aquitaine)
The defense will take place at 9h00 - Amphithéâtre RdC Broca Centre Broca Nouvelle-Aquitaine 146 Rue Léo Saignat, 33000 Bordeaux
in front of the jury composed of
- Mathieu BERANECK - Directeur de recherche - Integrative Neuroscience and Cognition Center - CNRS UMR 8002 CNRS, Université de Paris - Rapporteur
- Hélène BRAS - Chargée de recherche - Institut de Neurosciences de la Timone, UMR7289 - Rapporteur
- Agnès NADJAR - Professeure - Université de Bordeaux, INSERM, Neurocentre Magendie - Examinateur
- Pascal FOSSAT - Professeur - institut des maladies neurodégénératives, CNRS UMR 5293 Université de Bordeaux - Examinateur
In all vertebrates, the vestibular system detecs head movements to stabilize gaze and posture. This sensorimotor transformation mechanism takes place via the vestibulomotor pathways, which integrate vestibular sensory signals to elaborate a reflex motor command, vestibulo-ocular (VO) or vestibulo-spinal (VS), involved in oculomotor and postural control respectively. This transformation is specifically adapted to vestibulo-ocular and spinal constraints. Amphibians central vestibular neurons offer a unique model to understand the anatomo-functionnal maturation of vestibular networks in relation to the constraints and specificities of vestibular control of gaze and posture. The oculomotor system, the labyrinthine organs and the anatomy of the vestibular nuclei are not fundamentally modified during metamorphosis, whereas the postural-locomotor system and the vestibular sensory referential are totally reorganized. The aim of this thesis is to study the maturation and reorganization of vestibulomotor networks involved in vestibular reflex behaviors facing more or less significant changes during xenopus development. Intracellular recordings of larval central vestibular neurons, performed before metamorphosis, and in juveniles neurons, performed after metamorphosis, revealed two distinct electrophysiological phenotypes: phasic neurons with transient discharge, and tonic neurons with continuous discharge. Phasic neurons properties showed no major differences between larvae and juveniles. For tonic neurons, on the other hand, discharge dynamic appears reduced in juveniles. Proportion of phasic and tonic VS neurons changes radically from a majority of tonic neurons in larvae to a majority of phasic neurons in juveniles. Conversely, VO neurons remain predominantly tonic during metamorphosis. These different proportions of phasic and tonic neurons could be partly explained by specific neurogenic activity within the VS and VO populations during early larval development. These differences observed in the VO and VS neuronal groups underlie the greater or lesser remodeling of vestibulo-ocular and vestibulospinal reflex behaviors. Vestibulomotor networks are also reorganized during xenopus development. Another aspect of this thesis was to understand how the vestibulospinal pathways responsible for postural control of the axial system in the larva could adapt during metamorphosis to produce postural control in the juvenile. Neuronal tracing and spinal reflex recording revealed direct vestibular inputs to both rostral and caudal spinal motoneurons, respectively located spinal segments preserved in juveniles, and segments which degenerate during metamorphosis. These vestibulospinal pathways generate two types of postural activity, vestibulospinal reflexes and more or less continuous swimming activity. The presence of vestibulospinal pathways in spinal segments that will disappear in juveniles raises the question of how these pathways are reorganized during metamorphosis. It seems that part of larval VS neurons are maintained in juveniles, suggesting developmental plasticity mechanisms at the cellular level that adapt cellular properties and projections to dedicated vestibular constraints.
Study of the effects of sensorimotor experience on numerical processing
by Julie LENOIR (Institut de neurosciences cognitives et intégratives d'Aquitaine)
The defense will take place at 13h30 - Amphi 6 146 Rue Léo Saignat, 33000 Bordeaux
in front of the jury composed of
- Arnaud BADETS - Chargé de recherche - Université de Bordeaux - Directeur de these
- Maria Dolores DE HEVIA - Chargée de recherche - Université de Paris-Cité - Rapporteur
- Michael ANDRES - Professeur des universités - Université Catholique de Louvain - Rapporteur
- Sandrine MEJIAS - Maître de conférences - Université de Lille - Examinateur
- Thibaut BROUILLET - Professeur des universités - Université Paris-Nanterre - Examinateur
- David COHEN - Professeur des universités - praticien hospitalier - Sorbonne Université - Examinateur
From an embodied perspective of cognition, number processing and the spatial organization of motor responses mutually interact each other. The first experimental part of this thesis (Axis 1) examined how bodily experiences (i.e., hand movements) influence numerical cognition, revealing that sensorimotor experience affects this processing. In the context of numerical cognition grounded in the sensorimotor system, the two presented studies highlighted the constitutive role of the hand in understanding both continuous and discrete magnitudes. Study 1 demonstrated how observing and performing hand actions guided numerical processing, while Study 2 explored the effect of sensorimotor experiences on transcoding between analog codes (e.g., perceiving a distance) and symbolic codes (e.g., numbers). From the perspective of situated numerical cognition, the second part of this thesis (Axis 2) explored the spatial dimension of numerical cognition. The three contributions in this section showed that spatial perception and the integration of bodily reference frames influence the manipulation of numerical information in a reciprocal dynamic. Overall, the interactions between body, space, and numbers were shown to be adaptive, requiring an understanding of the spatial reference frames used and a precise theoretical framework for the observed phenomena.
Assessment of the effect of environmental factors on stilbenoids, consequences on their antimicrobial activities and search for protective solutions
by Anthony PEBARTHE-COURROUILH (Oenologie)
The defense will take place at 14h00 - ISVV, Amphithéâtre 210 Chemin de Leysotte, Institut des Sciences de la Vigne et du Vin (ISVV), CS50008, 33882 Villenave-d'Ornon
in front of the jury composed of
- Stéphanie CLUZET - Professeure - Université de Bordeaux - Directeur de these
- Muriel VIAUD - Directrice de recherche - Université Paris-Saclay - Rapporteur
- Virginie PUECH-PAGES - Maîtresse de conférences - Université Toulouse III - Paul Sabatier - Rapporteur
- Marie-France CORIO-COSTET - Directrice de recherche - INRAE - Examinateur
- Michel HERNOULD - Professeur - Université de Bordeaux - Examinateur
- Arnaud LANOUE - Maître de conférences - Université de Tours - Examinateur
Grapevine (Vitis vinifera L.), a crop of major economic importance throughout the world, is susceptible to numerous diseases that cause damage to the plant and its fruits, leading to losses in yield and wine quality. Downy mildew (Plasmopara viticola) and grey mold (Botrytis cinerea) are among the most damaging diseases. They are traditionally controlled by the use of synthetic or copper-based pesticides, which are both harmful to the environment and humans. For a more sustainable viticulture, it is necessary to develop alternative/complementary solutions, such as the use of biopesticides (natural antimicrobial compounds). In grapevine, stilbenoids, and more specifically stilbenes stricto sensu, are considered to be phytoalexins, as they display antimicrobial activities against numerous pathogens. These polyphenols can be induced following stresses, but are also constitutively accumulated in the plant, especially in woody parts such as canes which are co-products of viticulture. Extracts enriched in stilbenoids and obtained from such co-products, allowing them to be valorized, could be used as biopesticides. However, when applied in the vineyard, their antimicrobial activities are reduced, probably due to the stilbene instability under environmental conditions. Solar radiation and fungal enzymes have been identified as factors compromising the stability of certain stilbenoids. Therefore, the aim of this thesis was i) to study the stability of stilbenoids in a cane extract when exposed to UV-A (abiotic factor) and when in contact to B. cinerea laccases (biotic factor), ii) to assess the consequences of the action of these 2 factors on the antimicrobial potential of stilbenoids towards P. viticola and B. cinerea, and iii) to explore protection methods to ensure the stilbenoid stability. The results reveal that the majority of the stilbenoids in the extract, all stilbenes, were degraded by UV-A. Through a study on pure molecules including trans-ε-viniferin and trans-resveratrol, two major molecules of the cane extract, we noted that photo-transformations began with compound isomerization followed by oxidation leading to phenanthrenic structures. After UV-A exposure, the cane extract showed a drastic reduction in its antimicrobial activity against P. viticola and, to a lesser extent against B. cinerea. To limit the negative impact of UV-A, two approaches for photoprotecting stilbenes were studied: combination of the cane extract with flavonoids and emulsification of the extract. As a result, stilbenes were protected during the first minutes of exposure, in correlation with a partial maintenance of anti-mildew activity. Subsequently, the aim was to improve the stability of stilbenes in the presence of B. cinerea laccases, stilbene-oxidizing enzymes. Grape seed tannins, known as laccase inhibitors in wine, showed a greater capacity to inhibit laccase activity than stilbenoids. Among stilbenoids, only stilbenes s.s. were biotransformed by laccases, and to a lesser extend in the presence of tannins, in correlation with a slight gain in anti-mycelial activity. Finally, various tannin extracts were evaluated for their antimicrobial potential against downy mildew. The results highlight that grape-seed tannin extracts (condensed tannins) were the most inhibiting of oomycete development. In addition, the inhibitory activity of these extracts was also evaluated on sporangia release and zoospore mobility.
ED Sciences Physiques et de l'Ingénieur
Experimental studies of high-energy-density magnetized plasmas: in route to magneto-inertial fusion
by Christos VLACHOS (Centre Lasers Intenses et Applications)
The defense will take place at 14h30 - Amphithéâtre B, Bâtiment A29 351 cours de la Libération, 33405, Talence
in front of the jury composed of
- Joao Jorge SANTOS - Professor - Université de Bordeaux - Directeur de these
- Xavier RIBEYRE - Directeur de recherche - Commissariat à l'énergie atomique et aux énergies alternatives - Examinateur
- Olga ROSMEJ - Professor - Plasma Physics Department, GSI- Darmstadt - Rapporteur
- Michael TATARAKIS - Professor - Hellenic Mediterranean University - Directeur de these
- Fabrizio CONSOLI - Directeur de recherche - ENEA, Fusion and Technologies for Nuclear Safety and Security Department - Rapporteur
- Sophie BATON - Directrice de recherche - LULI, CNRS, CEA, École Polytechnique - Examinateur
- Ioannis FITILIS - Associate Professor - Hellenic Mediterranean University - Examinateur
This thesis is devoted to the exploration of magnetized high-energy-density (HED) plasmas driven by laser along with the understanding of the physics involved in the magnetic field (B-field) generation using laser-driven coil targets (LDCs). The latter is one of the possible strategies to magnetize HED plasmas in facilities without external pulsed power. One of the main axes was the dimensioning and the deployment of laser-driven cylindrical implosion experiments under seed B-fields to study magnetization effects on the implosions and, more specifically, on the transport of energy and magnetic flux in HED plasmas. The thesis presents K-shell emission X-ray spectra obtained from the radial implosion of Ar-doped D2-filled cylinders at the OMEGA laser, with 15 kJ of UV laser drive, that exhibit noticeable changes when a B-field is present, as compared to when it is not. This work was supported, through collaborations, by extended magnetohydrodynamics (MHD) and atomic physics simulations modeling the implosions, namely the transport of energy, magnetic flux and radiation. A multi-emission zone analysis of the experimental spectra allows for quantitative measurements of the plasma conditions in the core, indicating a 50% increase in temperature at half the mass density, when a 30 T seed field is applied. Besides, the predictions from atomic physics post-processing of extended-MHD simulations closely match the experimental spectra, providing compelling evidence that the core conditions at peak compression are in line with the impact of a 10 kT compressed field. The findings provide a foundation for verifying magnetized transport models in highly compressed plasmas – namely to test the microphysics involved in the transport coefficients used in the MHD macroscopic models – and to a more robust design of bigger scale experiments at LMJ and NIF. The other main axis of the thesis was the further understanding of seed B-field generation from LDCs. The thesis presents new measurements of such B-fields driven by high-energy laser pulses of sub-nanosecond or multiple nanosecond durations, either at IR or UV wavelengths, with typical intensities of the order of 1015 W/cm2. At the LULI2000 facility, the B-fields were characterized by proton deflectometry in two perpendicular probing axes. We succeeded to identify effects of both magnetic and electric fields on the probing protons, therefore quantifying both coil discharge current (of well identified polarity) and static charge at different times. Measured currents are broadly consistent with predictions from a laser-driven diode-current source and lumped RL circuit model, supporting the quasi-steady assessment of the discharges. Peak B-fields of 50 T at the center of 0.5 mm diameter coils were obtained at the explored, rather moderate, laser intensity. At the PALS facility, we obtained a more complete physical picture of the active current-driving region in LDCs, that is the laser interaction region. We simultaneously measured the B-field near the coil (by optical polarimetry) and the self-generated B-field in the laser-produced plasma plume (by complex interferometry), along with a thorough characterization of the plasma particle sources and ground discharges through the target holder. We discuss the correlation between the coil current and the plasma plume current across the plates, inferred respectively from the above B-field measurements. This platform allowed to optimize peak B-fields for different target geometrical parameters. Finally, we also discuss B-field measurements from LDCs driven by UV laser light at the OMEGA and LMJ facilities, as a prelude of their potential use for magnetizing large-scale HED plasma experiments. Different theoretical models for the evolution of LDC-generated B-fields are confronted with the experimental results of the various experiments.
ED Sociétés, Politique, Santé Publique
Gender socialisation in the family and at school and construction of relations to digital for girls and boys aged 7 to 10 : studies with children and teachers.
by Lisa FERICELLI (Laboratoire de Psychologie)
The defense will take place at 14h00 - Amphithéâtre Gintrac Bâtiment Q 3ter Pl. de la Victoire, 33000 Bordeaux
in front of the jury composed of
- Véronique ROUYER - Professeure des universités - Université de Bordeaux - Directeur de these
- Christine MORIN-MESSABEL - Professeure des universités - Université Lumière Lyon 2 - Rapporteur
- Cédric FLUCKIGER - Professeur des universités - Université de Lille - Rapporteur
- Isabelle COLLET - Professeure associée - Université de Genève - CoDirecteur de these
- Gaël PASQUIER - Maître de conférences - Université Paris Est Créteil - Examinateur
The increasing integration of digital technology into all sectors of society has redefined educational paradigms, highlighting the need to acquire skills in this field (MENJ, 2023). However, an imbalance exists in the engagement of women compared to men in digital fields and professions (Collet, 2019). Although numerous studies reveal gender inequalities in this field among adults, studies exploring these phenomena among children remain rare. Yet there are marked differences in the digital uses and career aspirations of adolescents (Coavoux, 2019; Fontar & Le Mentec, 2020; Monfort & Réguer-Petit, 2022). This research has revealed that gender stereotypes and social expectations play a significant role in shaping young people's technological skills and interests, thus influencing their future pathway. Therefore, it is relevant to explore the construction of people's relations to the digital from childhood onwards, since during this period, girls and boys are already having digital-related experiences in different contexts (Cordier, 2021; Dajez & Roucous, 2010). The models to which children are exposed in the family and at school are heterogeneous and can be conflicting. Children don't just passively receive these influences: they play an active role in their gender socialization and the appropriation of associated norms (Rouyer & Troupel, 2013). In this context, the aim of this research is to examine the construction of the relation to the digital (representations of the digital, the child's digital uses and interests, self-efficacy) of girls and boys aged 7 to 10, in the family and school contexts. To this end, we looked at the role of children's gender socialization in their living environments, notably the family and school, and the way in which children actively construct their relation with the digital world. With a particular focus on the school environment, we studied teachers' representations and practices in relation to digital technology, in order to document them and draw out perspectives for reflection on the teaching of digital technology in schools. More specifically, our research is divided into two main parts: firstly, a study to document the construction of children's relation to digital, based on individual interviews with 52 children and a questionnaire on family digital uses completed by parents (n=42). Secondly, a study focused on schools, including interviews with 8 teachers and observations of their practices during sessions using digital tools. Thematic analyses of children's interviews reveal that their digital uses are influenced by a variety of factors, such as access to digital tools, interest in these tools, their representations, particularly gendered ones, and their individual experiences in the family and at school. The varied experiences and points of view highlight the need for a nuanced approach to understanding children's digital relation, and for children's views on their experiences of socialization in relation to digital to be taken into account. Interviews with teachers show that their personal representations of the digital world seem to strongly influence their teaching practices. Classrooms observations during sessions involving digital tools revealed a variety of teaching methods, depending on the teacher and the digital tools at their disposal. Their analysis in terms of the “Equality web” (Collet et al., 2024) identified some teaching practices that do not foster an egalitarian environment during sessions involving digital tools.
Insights from mathematical models into COVID-19: Analyzing public health interventions and immunity dynamics
by Iris GANSER (Bordeaux Population Health Research Center)
The defense will take place at 14h00 - Amphitheâtre Louis Université de Bordeaux Bâtiment ISPED 146 rue Léo-Saignat 33076 BORDEAUX CEDEX
in front of the jury composed of
- Rodolphe THIEBAUT - Professeur des universités - Université de Bordeaux - Directeur de these
- Alison HILL - Assistant professor - Johns Hopkins University - Rapporteur
- Michael WHITE - Directeur de recherche - Institut Pasteur - Rapporteur
- Robert PLATT - Full professor - McGill University, Departments of Epidemiology, Biostatistics, and Occupational Health and of Pediatrics - Rapporteur
- David BUCKERIDGE - Full professor - McGill University, School of Population and Global Health - CoDirecteur de these
- Marta NUNES - Professeur à l'université - Centre International de Recherche en Infectiologie (CIRI) - Examinateur
- Temime LAURA - Professeur des universités - Conservatoire national des Arts et Métiers (Cnam) - Examinateur
The COVID-19 pandemic, caused by the SARS-CoV-2 virus, has led to significant morbidity and mortality, straining healthcare systems worldwide. Fundamental approaches for controlling viral spread and mitigating its impact are vaccinations and non-pharmaceutical interventions (NPIs). Before vaccines became available, governments relied on NPIs with largely unknown epidemiological and societal impacts. Despite numerous studies, the effectiveness of NPIs on COVID-19 dynamics remained uncertain, especially over multiple pandemic waves. With a gradual roll-out of vaccines, population immunity increased, but this increase in was counteracted by the emergence of immune-escaping variants of concern (VoCs) and waning of both infection- and vaccine-induced immunity. The long-term dynamics of this decline are currently not well characterized, particularly in the context of multiple infections and infections with different VoCs. Given the only partially observed nature of epidemics and their non-linear dynamics, mathematical models are uniquely suited for their analysis. In my thesis, I applied mathematical models to various COVID-19 data, from aggregated population-level data of infections and hospitalizations to antibody (Ab) titers in individuals, with the goal of quantifying the effectiveness of NPIs and vaccines, identifying protective Ab thresholds, and characterizing immunity waning dynamics. Specifically, my first objective was to estimate the effectiveness of NPIs and vaccines in France and explore counterfactual NPI and vaccine implementation scenarios. We developed a population-based mechanistic model, which we fit to epidemiological data in France from March 2020 to October 2021. The model showed a significant reduction in viral transmission by lockdowns, school closures, and curfews, though their effectiveness decreased over time. Simulations demonstrated that vaccines had saved nearly 160k lives over the study period, but an earlier implementation or a faster rollout could have prevented even more deaths. To understand why NPI effectiveness estimates vary across studies, we evaluated two methodologies in my second objective: mechanistic models and a commonly used two-step regression approach. The latter first estimates the reproductive number (Rt) and then regresses it against NPI parameters. Using simulated data of varying complexity, mechanistic models consistently showed minimal bias (0-5%) and high confidence interval (CI) coverage, whereas the two-step regressions had biases up to 20% and much lower CI coverage. The bias stemmed from the depletion of susceptibles and challenges in estimating Rt, indicating that caution is warranted with this method despite its simplicity and speed. Accurate epidemiological models require up-to-date parameters. My third objective was therefore two-fold: 1) to relate SARS-CoV-2 specific Ab levels to the risk of infection and 2) to characterize antibody waning. Using Ab data from over 220k Canadian blood donors between April 2020 and December 2023, we found that both anti-S and anti-N Abs reduced infection risk, with anti-N showing a stronger effect at lower titers. We used biphasic decay models to characterize waning dynamics and estimated that that 51.3% (95% CI 40.6-66.1%) of individuals would drop below detectable anti-N Ab levels within three years after a single infection. The duration of Ab detection increased after subsequent infections. However, antibodies waned within months below thresholds needed to attain substantial protection, even after multiple infections and vaccinations, indicating that continuous vaccine booster doses might be needed to sustain protection. The analyses I conducted in my PhD research highlight the importance of timely interventions and continuous monitoring of immunity to better prepare for future outbreaks. Moreover, I illustrated that mathematical models are a powerful tool to inform public health decision making and strategies.