Search

We optimize a general model of bioprocesses, which is nonconvex due to the microbial growth in the biochemical reactors. We formulate a convex relaxation and give conditions guaranteeing its exactness in both the transient and steady state cases. When the growth kinetics are modeled by the Monod function under constant biomass or the Contois function, the relaxation is a second-order cone program, which can be solved efficiently at large scales. We implement the model on a numerical example based on a wastewater treatment system.

International audience; This work showcases the performance of [NiFeSe] hydrogenase from Desulfomicrobium baculatum for solar-driven hydrogen generation in a variety of organic-based deep eutectic solvents. Despite its well-known sensitivity towards air and organic solvents, the hydrogenase shows remarkable performance under an aerobic atmosphere in these solvents when paired with a TiO2 photocatalyst. Tuning the water content further increases hydrogen evolution activity to a TOF of 60 +/- 3 s(-1) and quantum yield to 2.3 +/- 0.4 % under aerobic conditions, compared to a TOF of 4 s(-1) in a purely aqueous solvent. Contrary to common belief, this work therefore demonstrates that placing natural hydrogenases into non-natural environments can enhance their intrinsic activity beyond their natural performance, paving the way for full water splitting using hydrogenases.

Owing to their roles in the arsenic (As) biogeochemical cycle, microorganisms and plants offer significant potential for developing innovative biotechnological applications able to remediate As pollutions. This possible use in bioremediation processes and phytomanagement is based on their ability to catalyse various biotransformation reactions leading to, e.g. the precipitation, dissolution, and sequestration of As, stabilisation in the root zone and shoot As removal. On the one hand, genomic studies of microorganisms and their communities are useful in understanding their metabolic activities and their interaction with As. On the other hand, our knowledge of molecular mechanisms and fate of As in plants has been improved by laboratory and field experiments. Such studies pave new avenues for developing environmentally friendly bioprocessing options targeting As, which worldwide represents a major risk to many ecosystems and human health.

We studied the marine pollution of the sediments and coastal waters of the Cape Verde peninsula and the Petite Côte. The results show that the perception of the population is not always in line with scientific observations. The population reports a mainly moderate level of pollution, which has worsened over the last ten years. The majority of the sites studied do not comply with the microbiological quality of bathing water from a microbiological point of view. The quantities of microplastics at some sites are high. Macroplastics at sea are not significantly present at all the sites sampled. For TMEs only chromium and nickel are above the probable effect concentration (PEC) on marine organisms for the fraction ≤ 100 μm. The first trophic levels of the marine food chain do not seem to be affected by sediment toxicity, whereas the middle and upper levels are more affected. The majority (81%) of the sites studied revealed a level of ecotoxicity greater than 20%, via the Magallana gigas embryotoxicity test, which is known to be more sensitive than other bioassays. Consequently, the assessment of the toxicity of marine sediments shows that the majority of the selected sites appear to be in a poor ecotoxicological state. The inter-site variability of the results is due to the characteristics of the sites.; Nous avons étudié la pollution marine des sédiments et des eaux côtières de la presqu’ile du Cap-Vert et de la Petite côte. Les résultats montrent que la perception des populations n’est pas toujours en phase avec les observations scientifiques. La population signale un niveau de pollution principalement modérée mais qui s’est aggravé ces dix dernières années. La majorité des sites étudiés n’est pas conforme aux qualités microbiologiques de l’eau de baignade d’un point de vu microbiologique. Les quantités de microplastiques de certains sites sont élevées. Les macroplastiques en mer, ne sont pas significativement présents sur tous les sites échantillonnés. Pour les ETM seuls le chrome et le nickel sont au-dessus de la concentration d'effet probable (CEP) sur les organismes marins en ce qui concerne la fraction ≤ 100 μm. Les premiers niveaux trophiques de la chaine alimentaire du milieu marin ne semblent pas être affectés par la toxicité des sédiments alors que les niveaux moyens et supérieurs le sont davantage. La majorité (81%) des sites étudiés ont révélé un niveau d’écotoxicité supérieur à 20%, via le test d’embryotoxicité de Magallana gigas reconnu pour sa plus grande sensibilité par rapport aux autres bioessais. Par conséquent, l'évaluation de la toxicité des sédiments marins montre que la majorité des sites sélectionnés apparaissent dans un mauvais état écotoxicologique. La variabilité inter-sites des résultats sont dues aux caractéristiques des sites.

International audience; Storing data on DNA molecules is an alternative to traditional storage media by promoting density and longevity. This paper gives an overview of how data is stored on DNA and how the reading and writing processes (sequencing and synthesis) are currently implemented on semiconductor devices.

Sander Bekeschus and Sylvain Iséni equally contributed to this work as first authors.; International audience; Parameter studies of plasma treatment are informative about the optimal use of this technology in biomedical applications such as the argon radio-frequency plasma jet kINPen. However, the interdependence of the plasma-dissipated power in relation to input current and feed gas modulation on the resulting biological consequences has not been studied so far. To this end, a parameter study is presented, and the effect on human immune cell viability was investigated across different input current power and argon with oxygen/nitrogen feed gas admixture settings. It was found that with both nitrogen and oxygen admixtures, a concentration-dependent change in plasma-dissipated power emerged, which converged at 27.5 mA and 26.5 mA, respectively. The extent of cytotoxicity in immune cells confirmed the relevance of these findings, which were in congruency with the plasma dissipated powers identified. These findings underline the critical role and input parameter-dependent action of plasma sources for biomedical application.Dr Sander Bekeschus and Dr Sylvain Iséni equally contributed to this work as first authors.

International audience

International audience; Cell response heterogeneity upon treatment is a main obstacle in preclinical development of efficacious cancer drugs, due to the emergence of drug-tolerant cells. We have previously developed a single-cell workflow, Fate-Seq (1,2), to profile drug-tolerant persisters. Fate-Seq is based on the prediction of the drug response of each cells that are individually profile at the molecular level.To achieve this goal, Fate-Seq couples 3 single-cell techniques: first the prediction of the cell response phenotype (resistant or sensitive) for clonal cancer cells treated with a chosen drug, then the isolation of the predicted resistant cells from the predicted sensitive ones by laser-capture, and finally the RNAsequencing of each single-cell (sc-RNA-seq). These sc-RNA-seq dataset are then analyzed using random walks with restart, to prioritize the genes according to the drug-sensitive state of each cell and to identify the genes causing for cell drug-resistance (as opposed to the genes associated with drug-resistance, 3).To automatize and increase the prediction throughput, we present 3 new developments in our workflow using machine learning models to classify cell drug response, from the cell signaling dynamics observed with fate-seq, and to determine the molecular factors defining the drug efficacy of the cell type tested. Theses molecular factors represent good candidates to be targeted during a co-treatment, in combination with the first drug analyzed with our pipeline.We then introduce our eDRUGs (early Drug Response UpGraded) classifier, that combines mechanistic modeling of apoptosis (cell death) through cell signaling pathway, and machine learning classification models to predict cell drug response within an hour, using the fluorescent time-trajectories as input. This new method is twice as accurate as our previous prediction method (4).Finally, we will also propose a novel analysis method of sc-RNA-seq data obtained with Fate-Seq. This method consists in training binary classifiers on the scRNAseq expression data obtained .from the pipeline, using a range of models and explainable AI techniques such as DeepLift (5), in addition to clustering techniques, to obtain attribution scores for each gene. These scores are expected to reveal a reduced gene set, more manageable for drug combinations design.

Optogenetics arises as a valuable tool to precisely control genetic circuits in microbial cell factories. Light control holds the promise of optimizing bioproduction methods and maximize yields, but its implementation at different steps of the strain development process and at different culture scales remains challenging. In this study, we aim to control beta-carotene bioproduction using optogenetics inSaccharomyces cerevisiaeand investigate how its performance translates across culture scales. We built four lab-scale illumination devices, each handling different culture volumes, and each having specific illumination characteristics and cultivating conditions. We evaluated optogenetic activation and beta-carotene production across devices and optimized them both independently. Then, we combined optogenetic induction and beta-carotene production to make a light-inducible beta-carotene producer strain. This was achieved by placing the transcription of the bifunctional lycopene cyclase / phytoene synthase CrtYB under the control of the pC120 optogenetic promoter regulated by the EL222-VP16 light-activated transcription factor, while other carotenogenic enzymes (CrtI, CrtE, tHMG) were expressed constitutively. We show that illumination, culture volume and shaking impact differently optogenetic activation and beta-carotene production across devices. This enabled us to determine the best culture conditions to maximize light-induced beta-carotene production in each of the devices, reaching a content of up to 880 μg/gCDW. Our study exemplifies the stakes of scaling up optogenetics in devices of different lab scales and sheds light on the interplays and potential conflicts between optogenetic control and metabolic pathway efficiency. As a general principle, we propose that it is important to first optimize both components of the system independently, before combining them into optogenetic producing strains to avoid extensive troubleshooting. We anticipate that our results can help designing both strains and devices that could eventually lead to larger scale systems in an effort to bring optogenetics to the industrial scale.

International audience; Today, the community consensus to store information on DNA is to use short single strand DNA (ssDNA) molecules. This approach has somelimitations: encoding constraint, DNA stability, recovering DNA, sequencing technology, etc. To overcome them, we chose to store information onlong double-strand DNA (dsDNA) molecules. Our demonstration consists of storing the first articles of the declaration of human rights (4,2ko text document) on a single DNA molecule.

International audience; Cell response variability is a starting point in cancer drug resistance that has been difficult to analyze because the tolerant cell states are short lived. Here, we present fate-seq, an approach to isolate single cells in their transient states of drug sensitivity or tolerance before profiling. The drug response is predicted in live cells, which are laser-captured by microdissection before any drug-induced change can alter their states. This framework enables the identification of the cell-state signatures causing differential cell decisions upon treatment.

Considering an ever-growing global population, which hit 8 billion people in the fall of 2022, it is essential to find solutions to avoid the competition between human food and animal feed for croplands. Agricultural co -products have become important components of the circular economy with their use in animal feed. Their implementation was made possible by the addition of exogenous enzymes in the diet, especially carbohydrate -active enzymes (CAZymes). In this review, we describe the diversity and versatility of microbial CAZymes tar-geting non-starch polysaccharides to improve the nutritional potential of diets containing cereals and protein meals. We focused our attention on cellulases, hemicellulases, pectinases which were often found to be crucial in vivo. We also highlight the performance and health benefits brought by the exogenous addition of enzymatic cocktails containing CAZymes in the diets of monogastric animals. Taking the example of the well-studied commercial cocktail RovabioTM, we discuss the evolution, constraints and future challenges faced by feed en-zymes suppliers. We hope that this review will promote the use and development of enzyme solutions for in-dustries to sustainably feed humans in the future.

Effective and safe surgery results from a complex sociotechnical process prone to human error. Acquiring large amount of data on surgical care and modelling the process of surgery with artificial intelligence's computational methods could shed lights on system strengths and limitations and enable computer-based smart assistance. With this vision in mind, surgeons and computer scientists have joined forces in a novel discipline called Surgical Data Science. In this regard, operating room (OR) black boxes and surgical control towers are being developed to systematically capture comprehensive data on surgical procedures and to oversee and assist during operating rooms activities, respectively. Most of the early Surgical Data Science works have focused on understanding risks and resilience factors affecting surgical safety, the context and workflow of procedures, and team behaviors. These pioneering efforts in sensing and analyzing surgical activities, together with the advent of precise robotic actuators, bring surgery on the verge of a fourth revolution characterized by smart assistance in perceptual, cognitive and physical tasks. Barriers to implement this vision exist, but the surgical-technical partnerships set by ambitious efforts such as the OR black box and the surgical control tower are working to overcome these roadblocks and translate the vision and early works described in the manuscript into value for patients, surgeons and health systems.

Brain-computer interfaces (BCIs) transform neural signals into commands for effectors. They are mainly used as tools for functional compensation of impaired functions in disabled subjects. The Clinatec clinical trial “BCI and tetraplegia” aims at providing a proof of concept of long-term functional compensation of upper and lower limbs motor deficits in tetraplegics using a BCI. The clinical trial showed promising results for compensation of motor functions. However, the BCI’s usability remains to be improved. The decoder used to estimate the user’s intention from the processed neural signal (control decoder) must be fitted before the BCI can be used. This is done during dedicated training sessions, during which the user is directed to perform specific motor imagery tasks. Training sessions have to be held regularly in order to update the control decoder due to degradation of performances with time. This thesis proposes to limit the negative impact of training sessions. In this work, the control decoder is trained and updated using inferred labels instead of labels acquired during a dedicated training session. The labels are inferred using the output of the control decoder and neural correlates of task performance. In order to be usable in state-of-the-art motor BCIs, the adaptation process must be possible for control decoders with multiple discrete or continuous outputs. We argue that adaptation of a control decoder with multiple continuous outputs is best done using neural correlates of task performance that can be decoded continuously in time. Additionally, these neural correlates should be detected in the sensorimotor cortex due to the position of the implanted neural acquisition system. Using multiple ECoG datasets from a tetraplegic enrolled in the Clinatec BCI clinical trial, we show in this thesis that it is possible to detect such continuous in time neural correlates of task performance from the sensorimotor cortex, that control decoder labels can be inferred using these neural correlates, and finally that these labels can be used to successfully train decoders for discrete or continuous control. The subject used motor imagery to control a binary avatar (Runner MI dataset), a virtual exoskeleton with four discrete motor states (Exo dataset) or a hand-shaped cursor on a two-dimensional screen (Cursor dataset). In five-fold cross-validations, the mean and standard deviation of the area under the receiver operating characteristic curve (AUC of the ROC) for the decoder of neural correlates of task performance were 0.6225±0.0429 in the Runner MI dataset, 0.5677±0.0427 in the Exo dataset and 0.6570±0.0188 in the Cursor dataset. In a pseudo cross-validation simulating online use, the accuracy of the estimated labels was 64.9% in the Runner dataset and 64.5% in the Exo dataset. In the Cursor dataset, 63.3% of the estimated labels were less than 60° away from the actual labels. The AUC of the ROC of the control decoder was 0.6360±0.0958 in the Runner MI dataset when trained auto-adaptively compared to 0.8958±0.0153 when trained supervisedly, and to a chance level of 0.5007±0.0691. The multiclass generalization of the AUC of the ROC of the control decoder was 0.7595±0.0278 in the Exo dataset when trained auto-adaptively compared to 0.8177±0.0301 when trained supervisedly, and to a chance level of 0.5163±0.0580. In the Cursor dataset, the cosine similarity was 0.1589±0.0668 when trained auto-adaptively compared to 0.2107±0.0664 when trained in a classical supervised manner, and a chance level of -0.0231±0.0327. These results are promising for the future of auto-adaptive complex motor BCIs., Les Interfaces cerveau machine (ICMs) transforment les signaux neuronaux en commande pour des effecteurs. Elles sont principalement utilisées pour compenser les déficits d’handicapés moteurs. L’essai clinique « BCI et tétraplégie » de Clinatec a pour but the fournir une preuve de concept the compensation fonctionnelle chronique des déficits moteurs des quatre membres de tétraplégiques au moyen d’ICMs. Bien que les résultats obtenus soient prometteurs, l’usabilité des ICMs utilisées reste améliorable. Le décodeur utilisé pour transformer les signaux neuronaux en commandes (décodeur de contrôle) doit par exemple être entrainé avant de pouvoir utiliser l’ICM. Ceci est fait durant des séances d’entrainement spécifiques, durant lesquelles le patient doit effectuer des tâches d’imagerie motrices imposés. Ces séances d’entrainement constituent un temps mort et doivent être effectuées régulièrement afin de compenser les pertes de performances qui surviennent avec le temps. Cette thèse vise à limiter les effets négatifs des séances d’entrainements. Pour cela, le décodeur de contrôle est entrainé et mis à jour avec des labels estimés au lieu de labels obtenus lors de séances d’entrainement. Les labels sont estimés grâce à la sortie du décodeur de control ainsi qu’un décodeur de corrélats neuronaux de performance de tâche. Afin d’être utilisable avec les ICMs motrices les plus performantes, l’adaptation doit être faisable pour les décodeurs de contrôle qui ont plusieurs sorties continues. Nous suggérons qu’une telle adaptation requiert des corrélats neuronaux de performances de tâche acquis de manière continuelle dans le temps. Il est aussi nécessaire de détecter ces corrélats dans le cortex sensorimoteur à cause de la position des implants. Dans cette thèse, nous montrons sur plusieurs jeux de données ECoG d’un tétraplégique qu’il est possible de détecter continuellement des corrélats neuronaux de performance de tâche depuis le cortex sensorimoteur, que les labels des décodeurs de contrôle peuvent être estimés grâce à ces corrélats et enfin que ces labels peuvent être utilisés pour l’entrainement de décodeurs avec des sorties discrètes ou continues. Le patient a utilisé l’imagerie motrice pour contrôler un avatar binaire (jeu de données Runner MI), un exosquelette virtuel avec quatre états moteur discrets (jeu de données Exo) ou un curseur en deux dimensions continues sur un écran d’ordinateur (jeu de données Cursor). Lors de validations croisées à 5 blocs, la moyenne et la déviation standard de l’aire sous la courbe (AUC) ROC du décodeur de corrélats neuronaux de performance de tâche était de 0.6225±0.0429, 0.5677±0.0427 et 0.6570±0.0188 pour les jeux de données Runner MI, Exo et Cursor respectivement. Dans une simulation d’utilisation en ligne de de cette méthode d’adaptation, la précision des labels estimés était de 64.9% et 64.5% pour les jeux de données Runner MI et Exo. Pour le jeu de données Cursor, 63% des labels estimés étaient à moins de 60° d’écart des vrais labels. Dans cette simulation avec le jeu de donnée Runner MI, l’AUC de la ROC du décodeur de contrôle était de 0.6360±0.0958 quand entrainé de manière auto-adaptative, comparé à 0.8958±0.0153 pour un entrainement supervisé et à un niveau de chance de 0.5007±0.0691. Pour le jeu de données Exo, la généralisation multi classe de l’AUC de la ROC était 0.7595±0.0278 pour un entrainement auto-adaptative, comparé à 0.8177±0.0301 pour un entrainement supervisé et à un niveau de chance de 0.5163±0.0580. Pour le jeu de données Cursor, la similarité cosinus était 0.1589±0.0668 pour un entrainement auto-adaptative, comparé à 0.2107±0.0664 pour un entrainement supervisé et à un niveau de chance de -0.0231±0.0327. Ces résultats sont prometteurs pour le développement des ICMs motrices complexes auto-adaptatives.

Brain-computer interfaces (BCIs) transform neural signals into commands for effectors. They are mainly used as tools for functional compensation of impaired functions in disabled subjects. The Clinatec clinical trial “BCI and tetraplegia” aims at providing a proof of concept of long-term functional compensation of upper and lower limbs motor deficits in tetraplegics using a BCI. The clinical trial showed promising results for compensation of motor functions. However, the BCI’s usability remains to be improved. The decoder used to estimate the user’s intention from the processed neural signal (control decoder) must be fitted before the BCI can be used. This is done during dedicated training sessions, during which the user is directed to perform specific motor imagery tasks. Training sessions have to be held regularly in order to update the control decoder due to degradation of performances with time. This thesis proposes to limit the negative impact of training sessions. In this work, the control decoder is trained and updated using inferred labels instead of labels acquired during a dedicated training session. The labels are inferred using the output of the control decoder and neural correlates of task performance. In order to be usable in state-of-the-art motor BCIs, the adaptation process must be possible for control decoders with multiple discrete or continuous outputs. We argue that adaptation of a control decoder with multiple continuous outputs is best done using neural correlates of task performance that can be decoded continuously in time. Additionally, these neural correlates should be detected in the sensorimotor cortex due to the position of the implanted neural acquisition system. Using multiple ECoG datasets from a tetraplegic enrolled in the Clinatec BCI clinical trial, we show in this thesis that it is possible to detect such continuous in time neural correlates of task performance from the sensorimotor cortex, that control decoder labels can be inferred using these neural correlates, and finally that these labels can be used to successfully train decoders for discrete or continuous control. The subject used motor imagery to control a binary avatar (Runner MI dataset), a virtual exoskeleton with four discrete motor states (Exo dataset) or a hand-shaped cursor on a two-dimensional screen (Cursor dataset). In five-fold cross-validations, the mean and standard deviation of the area under the receiver operating characteristic curve (AUC of the ROC) for the decoder of neural correlates of task performance were 0.6225±0.0429 in the Runner MI dataset, 0.5677±0.0427 in the Exo dataset and 0.6570±0.0188 in the Cursor dataset. In a pseudo cross-validation simulating online use, the accuracy of the estimated labels was 64.9% in the Runner dataset and 64.5% in the Exo dataset. In the Cursor dataset, 63.3% of the estimated labels were less than 60° away from the actual labels. The AUC of the ROC of the control decoder was 0.6360±0.0958 in the Runner MI dataset when trained auto-adaptively compared to 0.8958±0.0153 when trained supervisedly, and to a chance level of 0.5007±0.0691. The multiclass generalization of the AUC of the ROC of the control decoder was 0.7595±0.0278 in the Exo dataset when trained auto-adaptively compared to 0.8177±0.0301 when trained supervisedly, and to a chance level of 0.5163±0.0580. In the Cursor dataset, the cosine similarity was 0.1589±0.0668 when trained auto-adaptively compared to 0.2107±0.0664 when trained in a classical supervised manner, and a chance level of -0.0231±0.0327. These results are promising for the future of auto-adaptive complex motor BCIs., Les Interfaces cerveau machine (ICMs) transforment les signaux neuronaux en commande pour des effecteurs. Elles sont principalement utilisées pour compenser les déficits d’handicapés moteurs. L’essai clinique « BCI et tétraplégie » de Clinatec a pour but the fournir une preuve de concept the compensation fonctionnelle chronique des déficits moteurs des quatre membres de tétraplégiques au moyen d’ICMs. Bien que les résultats obtenus soient prometteurs, l’usabilité des ICMs utilisées reste améliorable. Le décodeur utilisé pour transformer les signaux neuronaux en commandes (décodeur de contrôle) doit par exemple être entrainé avant de pouvoir utiliser l’ICM. Ceci est fait durant des séances d’entrainement spécifiques, durant lesquelles le patient doit effectuer des tâches d’imagerie motrices imposés. Ces séances d’entrainement constituent un temps mort et doivent être effectuées régulièrement afin de compenser les pertes de performances qui surviennent avec le temps. Cette thèse vise à limiter les effets négatifs des séances d’entrainements. Pour cela, le décodeur de contrôle est entrainé et mis à jour avec des labels estimés au lieu de labels obtenus lors de séances d’entrainement. Les labels sont estimés grâce à la sortie du décodeur de control ainsi qu’un décodeur de corrélats neuronaux de performance de tâche. Afin d’être utilisable avec les ICMs motrices les plus performantes, l’adaptation doit être faisable pour les décodeurs de contrôle qui ont plusieurs sorties continues. Nous suggérons qu’une telle adaptation requiert des corrélats neuronaux de performances de tâche acquis de manière continuelle dans le temps. Il est aussi nécessaire de détecter ces corrélats dans le cortex sensorimoteur à cause de la position des implants. Dans cette thèse, nous montrons sur plusieurs jeux de données ECoG d’un tétraplégique qu’il est possible de détecter continuellement des corrélats neuronaux de performance de tâche depuis le cortex sensorimoteur, que les labels des décodeurs de contrôle peuvent être estimés grâce à ces corrélats et enfin que ces labels peuvent être utilisés pour l’entrainement de décodeurs avec des sorties discrètes ou continues. Le patient a utilisé l’imagerie motrice pour contrôler un avatar binaire (jeu de données Runner MI), un exosquelette virtuel avec quatre états moteur discrets (jeu de données Exo) ou un curseur en deux dimensions continues sur un écran d’ordinateur (jeu de données Cursor). Lors de validations croisées à 5 blocs, la moyenne et la déviation standard de l’aire sous la courbe (AUC) ROC du décodeur de corrélats neuronaux de performance de tâche était de 0.6225±0.0429, 0.5677±0.0427 et 0.6570±0.0188 pour les jeux de données Runner MI, Exo et Cursor respectivement. Dans une simulation d’utilisation en ligne de de cette méthode d’adaptation, la précision des labels estimés était de 64.9% et 64.5% pour les jeux de données Runner MI et Exo. Pour le jeu de données Cursor, 63% des labels estimés étaient à moins de 60° d’écart des vrais labels. Dans cette simulation avec le jeu de donnée Runner MI, l’AUC de la ROC du décodeur de contrôle était de 0.6360±0.0958 quand entrainé de manière auto-adaptative, comparé à 0.8958±0.0153 pour un entrainement supervisé et à un niveau de chance de 0.5007±0.0691. Pour le jeu de données Exo, la généralisation multi classe de l’AUC de la ROC était 0.7595±0.0278 pour un entrainement auto-adaptative, comparé à 0.8177±0.0301 pour un entrainement supervisé et à un niveau de chance de 0.5163±0.0580. Pour le jeu de données Cursor, la similarité cosinus était 0.1589±0.0668 pour un entrainement auto-adaptative, comparé à 0.2107±0.0664 pour un entrainement supervisé et à un niveau de chance de -0.0231±0.0327. Ces résultats sont prometteurs pour le développement des ICMs motrices complexes auto-adaptatives.

International audience

Guanine quadruplex nucleic acids (G4s) are involved in key biological processes such as replication or transcription. Beyond their biological relevance, G4s find applications as biotechnological tools since they readily bind hemin and enhance its peroxidase activity, creating a G4-DNAzyme. The biocatalytic properties of G4-DNAzymes have been thoroughly studied and used for biosensing purposes. Despite hundreds of applications and massive experimental efforts, the atomistic details of the reaction mechanism remain unclear. To help select between the different hypotheses currently under investigation, we use extended explicit-solvent molecular dynamics (MD) simulations to scrutinize the G4/hemin interaction. We find that besides the dominant conformation in which hemin is stacked atop the external G-quartets, hemin can also transiently bind to the loops and be brought to the external G-quartets through diverse delivery mechanisms. The simulations do not support the catalytic mechanism relying on a wobbling guanine. Similarly, catalytic role of the iron-bound water molecule is not in line with our results, however, given the simulation limitations, this observation should be considered with some caution. The simulations rather suggest tentative mechanisms in which the external G-quartet itself could be responsible for the unique H2O2-promoted biocatalytic properties of the G4/hemin complexes. Once stacked atop a terminal G-quartet, hemin rotates about its vertical axis while readily sampling shifted geometries where the iron transiently contacts oxygen atoms of the adjacent G-quartet. This dynamics is not apparent from the ensemble-averaged structure. We also visualize transient interactions between the stacked hemin and the G4 loops. Finally, we investigated interactions between hemin and on-pathway folding intermediates of the parallel-stranded G4 fold. The simulations suggest that hemin drives the folding of parallel-stranded G4s from slip-stranded intermediates, acting as a G4 chaperone. Limitations of the MD technique are briefly discussed.For Table of Contents Only

The continuing efforts to exploit the death receptor agonists, such as the tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), for cancer therapy, have largely been impaired by the anti-apoptotic and pro-survival signalling pathways leading to drug resistance. Cell migration, invasion, differentiation, immune evasion and anoikis resistance are plastic processes sharing features of the epithelial-to-mesenchymal transition (EMT) that have been shown to give cancer cells the ability to escape cell death upon cytotoxic treatments. EMT has recently been suggested to drive a heterogeneous cellular environment that appears favourable for tumour progression. Recent studies have highlighted a link between EMT and cell sensitivity to TRAIL, whereas others have highlighted their effects on the induction of EMT. This review aims to explore the molecular mechanisms by which death signals can elicit an increase in response heterogeneity in the metastasis context, and to evaluate the impact of these processes on cell responses to cancer therapeutics.

International audience; ► Context – Actual storage technologies (flash memory, hard drives,magnetic tapes,..) are outpaced by the exponential rise of digital dataproduction ► Advantages of DNA storage [2] – Density of 1021 bytes in one gram(106 times more compact than hard disks), durability for centuries,energy cost close to zero (molecules kept at room temperature with nomaintenance)► Motivation – Introducing security to ensure the confidentiality of thedata stored into DNA molecules

National audience

International audience; N-acetylglucosamine containing compounds acting as pathogenic or symbiotic signals are perceived by plant-specific Lysin Motif Receptor-Like Kinases (LysM-RLKs). The molecular mechanisms of this perception are not fully understood, notably those of lipo-chitooligosaccharides (LCOs) produced during root endosymbioses with nitrogen-fixing bacteria or arbuscular mycorrhizal fungi. In Medicago truncatula, we previously identified the LysM-RLK LYR3 (MtLYR3) as a specific LCO-binding protein. We also showed that the absence of LCO binding to LYR3 of the non-mycorrhizal Lupinus angustifolius, (LanLYR3), was related to LysM3, which differs from that of MtLYR3 by several amino acids and, particularly, by a critical tyrosine residue absent in LanLYR3. Here, we aimed to define the LCO binding site of MtLYR3 by using molecular modelling and simulation approaches, combined with site-directed mutagenesis and LCO binding experiments. 3D models of MtLYR3 and LanLYR3 ectodomains were built, and homology modelling and molecular dynamics (MD) simulations were performed. Molecular docking and MD simulation on the LysM3 identified potential key residues for LCO binding. We highlighted by steered MD simulations that in addition to the critical tyrosine, two other residues were important for LCO binding in MtLYR3. Substitution of these residues in LanLYR3-LysM3 by those of MtLYR3-LysM3 allowed the recovery of high-affinity LCO binding in experimental radioligand-binding assays. An analysis of selective constraints revealed that the critical tyrosine has experienced positive selection pressure and is absent in some LYR3 proteins. These findings now pave the way to uncover the functional significance of this specific evolutionary pattern.

L'institution a financé les frais de publication pour que cet article soit en libre accès; International audience; Objective: Between-session non-stationarity is a major challenge of current Brain-Computer Interfaces (BCIs) that affects system performance. In this paper, we investigate the use of channel selection for reducing between-session non-stationarity with Riemannian BCI classifiers. We use the Riemannian geometry framework of covariance matrices due to its robustness and promising performances. Current Riemannian channel selection methods do not consider between-session non-stationarity and are usually tested on a single session. Here, we propose a new channel selection approach that specifically considers non-stationarity effects and is assessed on multi-session BCI data sets. Methods: We remove the least significant channels using a sequential floating backward selection search strategy. Our contributions include: 1) quantifying the non-stationarity effects on brain activity in multi-class problems by different criteria in a Riemannian framework and 2) a method to predict whether BCI performance can improve using channel selection. Results: We evaluate the proposed approaches on three multi-session and multi-class mental tasks (MT)-based BCI datasets. They could lead to significant improvements in performance as compared to using all channels for datasets affected by between-session non-stationarity and to significant superiority to the state-of-the-art Riemannian channel selection methods over all datasets, notably when selecting small channel set sizes. Conclusion: Reducing non-stationarity by channel selection could significantly improve Riemannian BCI classification accuracy. Significance: Our proposed channel selection approach contributes to make Riemannian BCI classifiers more robust to between-session non-stationarities.

As is the case in several research domains, data sharing is still scarce in the field of Brain-Computer Interfaces (BCI), and particularly in that of passive BCIs—i.e., systems that enable implicit interaction or task adaptation based on a user's mental state(s) estimated from brain measures. Moreover, research in this field is currently hindered by a major challenge, which is tackling brain signal variability such as cross-session variability. Hence, with a view to develop good research practices in this field and to enable the whole community to join forces in working on cross-session estimation, we created the first passive brain-computer interface competition on cross-session workload estimation. This competition was part of the 3rd International Neuroergonomics conference. The data were electroencephalographic recordings acquired from 15 volunteers (6 females; average 25 y.o.) who performed 3 sessions—separated by 7 days—of the Multi-Attribute Task Battery-II (MATB-II) with 3 levels of difficulty per session (pseudo-randomized order). The data -training and testing sets—were made publicly available on Zenodo along with Matlab and Python toy code (https://doi.org/10.5281/zenodo.5055046). To this day, the database was downloaded more than 900 times (unique downloads of all version on the 10th of December 2021: 911). Eleven teams from 3 continents (31 participants) submitted their work. The best achieving processing pipelines included a Riemannian geometry-based method. Although better than the adjusted chance level (38% with an α at 0.05 for a 3-class classification problem), the results still remained under 60% of accuracy. These results clearly underline the real challenge that is cross-session estimation. Moreover, they confirmed once more the robustness and effectiveness of Riemannian methods for BCI. On the contrary, chance level results were obtained by one third of the methods—4 teams- based on Deep Learning. These methods have not demonstrated superior results in this contest compared to traditional methods, which may be due to severe overfitting. Yet this competition is the first step toward a joint effort to tackle BCI variability and to promote good research practices including reproducibility.

International audience; The dnarXiv projects 1 aims to explore various strategies for DNA storage. We have designed an experimental DNA storage platform allowing both to conduct real and/or in-silico experiments. The platform includes different modules generally used in the write/read DNA storage process: encoding, synthesis, molecule design, storage, molecule selection, genomic data processing, decoding. This is a flexible environment for testing various approaches simply by substituting new modules to the existing ones.

International audience; envisioned in 2025. To respond to associated storage needs, technologies as flash memoryand hard drives reach their limits in terms of: density; energy and ecological costs; durabilityfor long term storage. In this context, data storage on DNA molecules has recently been shownas very promising. DNA storage could be 106 times more dense than hard drives, with alifetime 10 to 20 times longer and an energy consumption closed to zero (molecules can bekept at room temperature with no maintenance).In this work, we are interested in securing archived data. DNA storage being a new technology,the opportunity presents itself to integrate this critical aspect at the biological level, contrarilyto what has been done for electronical storage means. In fact, information must be securedat every step of the DNA data storage chain. Data integrity and confidentiality are among themain issues with threats like data modification (e.g. writing of new data) or the theft of theDNA storage support by an attacker. Herein, we propose a solution for writing encrypted dataonto synthetic DNA molecules considering DNA synthesis and the error-correction codeconstraints. Indeed, DNA sequences should conform to structural constraints dictated by thisbiological process and sequencing.

International audience; Miniprotein binders hold a great interest as a class of drugs that bridges the gap between monoclonal antibodies and small molecule drugs. Like monoclonal antibodies, they can be designed to bind to therapeutic targets with high affinity, but they are more stable and easier to produce and to administer. In this chapter, we present a structure-based computational generic approach for miniprotein inhibitor design. Specifically, we describe step-by-step the implementation of the approach for the design of miniprotein binders against the SARS-CoV-2 coronavirus, using available structural data on the SARS-CoV-2 spike receptor binding domain (RBD) in interaction with its native target, the human receptor ACE2. Structural data being increasingly accessible around many protein-protein interaction systems, this method might be applied to the design of miniprotein binders against numerous therapeutic targets. The computational pipeline exploits provable and deterministic artificial intelligence-based protein design methods, with some recent additions in terms of binding energy estimation, multistate design and diverse library generation.

International audience; The electro-fermentation is a novel process that consists in coupling a microbial fermentative metabolism with an electrochemical system. In such a process the electrodes act either as electron sinks or sources modifying the fermentation balance of a microbial fermentative metabolism and provide new options for the control of microbial activity. A theoretical framework for the analysis and control of fermentations using electro-fermentation is currently lacking. In this paper, we propose a simple electro-fermentation model in which a population of fermentative bacteria switch between two metabolic behaviors in response to different electrode potentials. We then mathematically analyze optimal strategies to maximize the production of one of the rising products in a batch fermentation using Pontryagin's Maximum Principle. The obtained results show that, in some experimental configurations, a dynamic control of the electrode potential is required for the maximization of the desired product. Consequences of the obtained optimal strategy for driving electro-fermentation experiments are discussed through a realistic example. This analysis also highlights that the transition rates between fermentation and electro-fermentation behaviors are currently unknown and would be crucial to quantify in order to apply such a control approach.

International audience; The periodically forced light-limited Droop model represents microalgae growth under co-limitation by light and a single substrate, accounting for periodic fluctuations of factors such as light and temperature. In this paper, we describe the global dynamics of this model, considering general monotone growth and uptake rate functions. Our main result gives necessary and sufficient conditions for the existence of a positive periodic solution i.e. a periodic solution characterized by the presence of microalgae) which is globally attractive. In our approach, we reduce the model to a cooperative planar periodic system. Using results on periodic Kolmogorov equations and on monotone sub-homogeneous dynamical systems, we describe the global dynamics of the reduced system. Then, using the theory of asymptotically periodic semiflows, we extend the results on the reduced system to the original model. To illustrate the applicability of the main result, we include an example considering a standard microalgae population model.

International audience; Because of their innocuity, Au nanowires present an interesting field of applications in biology and, particularly, in cancer therapy. Since their morphology and distribution can greatly affect their performances, being able to control their mode of growth is important. Various elaboration techniques including “top-down” and “bottom-up” approaches have been developed. In this work, we propose an efficient maskless method to grow Au nanowires with the help of hydrogen plasma treatment of Au thin films. We have been able to grow Au nanowires by taking advantage of spinodal dewetting of an Au thin film and the supply of silicon radicals resulting from hydrogen plasma etching of amorphous silicon coating the walls of the reactor. A variety of techniques have been applied to study the microstructure and the optical properties of Au nanowires. A strong photothermal effect of Au nanowires has been demonstrated with the help of visible laser light. In order to study the nanowire growth, the transport of Au atoms is discussed, and a growth mechanism is proposed.

Background It is now widespread in livestock and plant breeding to use genotyping data to predict phenotypes with genomic prediction models. In parallel, genomic annotations related to a variety of traits are increasing in number and granularity, providing valuable insight into potentially important positions in the genome. The BayesRC model integrates this prior biological information by factorizing the genome according to disjoint annotation categories, in some cases enabling improved prediction of heritable traits. However, BayesRC is not adapted to cases where markers may have multiple annotations. Results We propose two novel Bayesian approaches to account for multi-annotated markers through a cumulative (BayesRC+) or preferential (BayesRC$$\pi$$ π ) model of the contribution of multiple annotation categories. We illustrate their performance on simulated data with various genetic architectures and types of annotations. We also explore their use on data from a backcross population of growing pigs in conjunction with annotations constructed using the PigQTLdb. In both simulated and real data, we observed a modest improvement in prediction quality with our models when used with informative annotations. In addition, our results show that BayesRC+ successfully prioritizes multi-annotated markers according to their posterior variance, while BayesRC$$\pi$$ π provides a useful interpretation of informative annotations for multi-annotated markers. Finally, we explore several strategies for constructing annotations from a public database, highlighting the importance of careful consideration of this step. Conclusion When used with annotations that are relevant to the trait under study, BayesRC$$\pi$$ π and BayesRC+ allow for improved prediction and prioritization of multi-annotated markers, and can provide useful biological insight into the genetic architecture of traits.

Background:Extracellular matrix (ECM) plays a pivotal role in many physiological processes. ECM macromolecules and associated factors differ according to tissues, impact cell differentiation, and tissue homeostasis. Dental pulp ECM may differ from other oral tissues and impact mineralization. Thus, the present study aimed to identify the matrisome of ECM proteins derived from human dental pulp stem cells (DPSCs) and its ability to regulate mineralization even in cells which do not respond to assaults by mineralization, the human gingival fibroblasts (GF).Methods:ECM were extracted from DPSCs cultured in normal growth medium supplemented with L-ascorbic acid (N-ECM) or in osteogenic induction medium (OM-ECM). ECM decellularization (dECM) was performed using 0.5% triton X-100 in 20 mM ammonium hydroxide after 21 days. Mass spectrometry and proteomic analysis identified and quantified matrisome proteins.Results:The dECM contained ECM proteins but lacked cellular components and mineralization. Interestingly, collagens (COL6A1, COL6A2, and COL6A3) and elastic fibers (FBN1, FBLN2, FN1, and HSPG2) were significantly represented in N-ECM, while annexins (ANXA1, ANXA4, ANXA5, ANXA6, ANXA7, and ANXA11) were significantly overdetected in OM-ECM. GF were reseeded on N-dECM and OM-dECM and cultured in normal or osteogenic medium. GF were able to attach and proliferate on N-dECM and OM-dECM. Both dECM enhanced mineralization of GF at day 14 compared to tissue culture plate (TCP). In addition, OM-dECM promoted higher mineralization of GF than N-dECM although cultured in growth medium.Conclusions:ECM derived from DPSCs proved to be osteoinductive, and this knowledge supported cell-derived ECM can be further utilized for tissue engineering of mineralized tissues.

International audience; Cell response heterogeneity upon drug treatment is a leading cause of reduced drug efficacy in preclinical cancer research. Although single-cell studies have revealed the depth of cellular heterogeneity observed between in tumor cells, the regulatory impact of cell variability on therapeutic response remains unclear. Here, we present a new ODE model of the extrinsic apoptosis initiation by death-ligands. This model is calibrated on fluorescent time-trajectories (FRET) of hundreds of clonal HeLa cells treated with different amounts of Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL). By highlighting the different steps in the regulation of apoptosis, and the associated timeline, we locate an initial cell fate decision just after TRAIL binding and the presence of additional regulation at the receptor that only benefits the drug-sensitive population. Then, our study provides evidence that increasing the dose of TRAIL actually has small effects within each population (resistant or sensitive) but rather accentuates the differences between the two, affecting the population dynamics in two different ways depending on their response to the drug. Finally, the distribution of 3 parameters of our mechanistic model, according to the cell drug response, suggests the existence of an determinant threshold in C8 dynamics, independent of the drug dose, that distinguishes cells drug-resistant or sensitive, that could be used to control or predict cell drug-response in the future.

International audience; User-specific settings are known to enhance Brain-Computer Interface (BCI) performances. In particular, the optimal frequency band and time segment for oscillatory activity classification are highly user-dependent and many selection methods have been developed in the past two decades. However, it has not been studied well whether those conventional methods can provide optimal settings for the Riemannian BCIs, a recent family of BCI systems that utilize different data representation, based on covariance matrices, compared to conventional BCI pipelines. In this work, we proposed a novel frequency band and time segment selection method considering class distinctiveness on the Riemannian manifold. The class distinctiveness of each combination of frequency band and time segment is quantified based on inter-class distance and intra-class variance on the Riemannian manifold. An advantage of this method is the user-specific settings can be adjusted without computationally heavy optimization steps. To the best of our knowledge, this is the first optimization method for selecting both the frequency band and the time segment on the Riemannian manifold. We evaluated the contributions of the 3 different selection models (frequency band, time or frequency band+time), comparing classification accuracy with a baseline (a fixed frequency band of 8-30 Hz and a fixed 2s time window) and a conventional popular method for non-Riemannian BCIs, on the BCI competition IV dataset 2a (2-class motor imagery). Our method showed higher average accuracy than baseline and a conventional method in all three models, and especially the frequency band selection model showed the highest performance. This preliminary result suggests the importance of developing new selection algorithms considering the properties of the manifold, rather than directly applying methods developed prior to the rise of Riemannian BCIs as they are.

Industry and health care are demanding rapid and inexpensive means for microbial identification. Thanks to its low cost and practical advantages, Petri dish culture is a ubiquitous tool in microbiology, but the sole observation of microbial colonies does not offer a reliable diagnosis. Identification in itself depends on secondary analysis, such as chemical reactions, PCR or mass spectrometry, which require specific sample preparation, which involves additional costs and delays. That is why several dish imaging systems have already been tested to automate the observation of cultures and to propose an identification directly on the Petri dish. However, these systems are generally limited to the visible and near infrared range (400 - 1000 nm), which only provides information on the morphotype of the microorganism colonies and therefore limits the identification accuracy.This thesis focuses on the development of a multispectral imaging system in the mid-infrared. In this wavelength range, images provide information on both the morphotype and the chemical composition of the observed colonies. This non-destructive and label-free imaging could provide species identification of colonies grown on agar, while opening the way to new applications. An experimental system, combining quantum cascade lasers as a light source and a microbolometers array as an imager, allowed the acquisition of images of colonies at nine wavelengths between 5 and 8 µm. 2253 colonies belonging to eight species of common microorganisms were imaged. For one of the species, Staphylococcus epidermidis, three different strains were analyzed to test the typing capabilities of the system.After acquisition, several image classification methods were tested to obtain an average correct identification rate of 94.4%., L’industrie et le domaine de la santé sont demandeurs de moyens rapides et peu couteux pour l’identification de microorganismes. Des avantages pratiques et un coût réduit font de la culture en boite de Petri un outil omniprésent en microbiologie, mais l’observation des colonies microbiennes n’offre pas de diagnostic fiable sans validation par une analyse secondaire. Ces analyses, qu’il s’agisse de réactions chimiques, d’une PCR ou de spectrométrie de masse, demandent une préparation d’échantillon spécifique impliquant des coûts et des délais supplémentaires. C’est pourquoi plusieurs systèmes d’imagerie de boite ont déjà été expérimentés pour automatiser l’observation des cultures et proposer une identification directement sur la boite de Petri. Toutefois, ceux-ci se cantonnent jusqu’à présent au domaine du visible et du proche infrarouge (400 – 1000 nm), ce qui n’apporte des informations que sur le morphotype des colonies de micro-organismes et limite la précision d’identification.Cette thèse est consacrée à l’exploration d’un système d’imagerie multispectrale dans l’infrarouge moyen. Cette gamme de longueurs d’onde apporte des informations à la fois sur le morphotype et sur la composition chimique des colonies observées. Cette imagerie non destructive et sans marquage pourrait fournir une identification à l’espèce des colonies cultivées sur agar, tout en ouvrant la voie à de nouvelles applications. Un système expérimental, combinant des lasers à cascade quantique comme source lumineuse et une matrice de microbolomètres comme imageur, a permis l’acquisition d’images de colonies à neuf longueurs d’ondes comprises entre 5 et 8 µm. 2253 colonies appartenant à huit espèces de microorganismes communs ont ainsi été imagées. Pour l’une des espèces, Staphylococcus epidermidis, trois souches différentes ont été analysées pour tester la capacité du système en matière de typage.Après acquisition, plusieurs méthodes de classification d’image ont été testées pour aboutir à un taux d’identification correct moyen de 94.4 %.

International audience

The abstract is available here: https://uscholar.univie.ac.at/o:1674457

In his prescient article “More is different”, P. W. Anderson counters the reductionist argument by highlighting the crucial role of emergent properties in science. This is particularly true in biology, where complex macroscopic behaviours stem from communication and interaction loops between much simpler elements. As an illustration, I hereby present three different instances in which I developed and used quantitative methods in order to learn new biological processes.For instance, the regulation and eventual rejection of tumours by the immune system is the result of multiple positive and negative regulation networks, influencing both the behaviour of the cancerous and immune cells. To mimic these complex effects in-vitro, I designed a microfluidic assay to challenge melanoma tumour spheroids with multiple T cells and observe the resulting interactions with high spatiotemporal resolution over long (>24h) periods of time. Using advanced image analysis combined with mathematical modelling I demonstrate that a positive feedback loop drives T cell accumulation to the tumour site, leading to enhanced spheroid fragmentation. This study sheds light on the initiation if the immune response at the single cell scale: showing that even the very first contact between T cell and tumour spheroid increases the probability of the next T cell to come to the tumour. It also shows that it is possible to recapitulate complex antagonistic behaviours in-vitro, which paves the way for the elaboration of more sophisticated protocols, involving for example a more complex tumour micro-environment.Many biological processes are the result of complex interactions between cell types, particularly so during development. The foetal liver is the locus of the maturation and expansion of the hematopoietic system, yet little is known about its structure and organisation. New experimental protocols have been recently developed to image this organ and I developed tools to interpret and quantify these data, enabling the construction of a “network twin” of each foetal liver. This method makes it possible to combine the single-cell scale and the organ scale in the analysis, revealing the accumulation of myeloid cells around the blood vessels irrigating the foetal liver at the final stages of organ development. In the future, this technique will make it possible to analyse precisely the environmental niches of cell types of interest in a quantitative manner. This in turn could help us understand the developmental steps of crucial cell types such as hematopoietic stem cells.The interactions between bacteria and their environment is key to understanding the emergence of complex collective behaviours such a biofilm formation. One mechanism of interest is that of rheotaxis, whereby bacterial motion is driven by gradients in the shear stress of the fluid the cells are moving in. I developed a framework to calculate the semi-analytical equations guiding bacteria movement in shear stress. These equations predict behaviours that aren’t observed experimentally, but the discrepancy is solved once rotational diffusion is taken into account. Experimental results are well-fitted by the theoretical prediction: bacteria in droplets segregate asymmetrically when a shear is generated in the media.Although relating to very different topics, these three studies highlight the pertinence of quantitative approaches for understanding complex biological phenomena: biological systems are more than the sum of their constituents.a; Les systèmes biologiques sont bien plus que la somme de leurs constituants. En effet, ils sont souvent caractérisés par des comportements macroscopiques complexes résultant de boucles d'interactions et de rétroactions. Par exemple, la régulation et le rejet éventuel des tumeurs par le système immunitaire est le résultat de multiples réseaux de régulation, influençant à la fois le comportement des cellules cancéreuses et immunitaires. Pour simuler ces effets complexes in-vitro, j'ai conçu une puce microfluidique permettant de confronter des sphéroïdes de mélanome à de multiples cellules T et d'observer les interactions qui en résultent avec une haute résolution spatio-temporelle et sur de longues périodes de temps. En utilisant de l'analyse d'images avancée, combinée à des modèles mathématiques, je démontre qu'une boucle de rétroaction positive conduit l'accumulation de cellules T sur la tumeur, ayant pour conséquence une fragmentation accrue des sphéroïdes. Cette étude met en lumière l'initiation de la réponse immunitaire à l'échelle de la cellule unique : elle montre que même le tout premier contact entre une cellule T et un sphéroïde tumoral augmente la probabilité que la cellule T suivante arrive sur la tumeur. Elle montre également qu'il est possible de récapituler des comportements antagonistes complexes in-vitro, ce qui ouvre la voie à l'élaboration de protocoles plus sophistiqués, impliquant par exemple un micro-environnement tumoral plus complexe.De nombreux processus biologiques sont le résultat d'interactions entre de multiples types de cellules, en particulier au cours du développement. Le foie fœtal est le lieu de la maturation et de l'expansion du système hématopoïétique, mais on sait peu de choses sur sa structure et son organisation. De nouveaux protocoles expérimentaux ont été récemment mis au point pour imager cet organe et j'ai développé des outils pour interpréter et quantifier ces données, permettant la construction d'un "réseau jumeau" de chaque foie fœtal. Cette méthode permet de combiner les échelles unicellulaire et de l'organe dans une seule analyse, révélant l'accumulation de cellules myéloïdes autour des vaisseaux sanguins irriguant le foie fœtal aux derniers stades du développement de l'organe. À l'avenir, cette technique permettra d'analyser précisément les environnements de cellules d'intérêt de manière quantitative. Ceci pourrait à son tour nous aider à comprendre les étapes du développement de types cellulaires cruciaux tels que les cellules souches hématopoïétiques.Les interactions entre les bactéries et leur environnement sont essentielles pour comprendre l'émergence de comportements collectifs complexes tels que la formation de biofilms. Un mécanisme d'intérêt est celui de la rhéotaxie, par lequel le mouvement bactérien est entraîné par les gradients de la contrainte de cisaillement du fluide dans lequel les cellules se déplacent. J'ai développé une méthode pour calculer les équations semi-analytiques guidant le mouvement des bactéries dans la contrainte de cisaillement. Ces équations prédisent des comportements qui ne sont pas observés expérimentalement, mais la divergence est résolue une fois que la diffusion rotationnelle est prise en compte. Les résultats expérimentaux correspondent bien à la prédiction théorique : les bactéries dans les gouttelettes se séparent de manière asymétrique lorsqu'un cisaillement est généré dans le milieu.

International audience; A limiting factor towards the wide real-life use of wearables devices for continuous healthcare monitoring is their cumbersome and obtrusive nature. This is particularly true for electroencephalography (EEG) recordings, which require the placement of multiple electrodes in contact with the scalp. In this work we propose to identify the optimal wearable EEG electrode setup , in terms of minimal number of electrodes, comfortable location and performance, for EEG-based event detection and monitoring. By relying on the demonstrated power of autoencoder (AE) networks to learn latent representations from high-dimensional data, our proposed strategy trains an AE architecture in a one-class classification setup with different electrode setups as input data. The model performance is assessed using the F-score. Alpha waves detection is the use case, through which we demonstrate that the proposed method allows to detect an alpha state from an optimal setup. The socalled wearable configuration, consisting of electrodes in the forehead and behind the ear, is the chosen optimal setup , with an average F-score of 0.78. Our results suggest that a learning-based approach can be used to enable the design and implementation of optimized wearable devices for real-life event related healthcare monitoring.

International audience; Regulation of RNA abundance and localization is a key step in gene expression control. Single-molecule RNA fluorescence in situ hybridization (smFISH) is a widely used single-cell-single-molecule imaging technique enabling quantitative studies of gene expression and its regulatory mechanisms. Today, these methods are applicable at a large scale, which in turn come with a need for adequate tools for data analysis and exploration. Here, we present FISH-quant v2, a highly modular tool accessible for both experts and non-experts. Our user-friendly package allows the user to segment nuclei and cells, detect isolated RNAs, decompose dense RNA clusters, quantify RNA localization patterns and visualize these results both at the single-cell level and variations within the cell population. This tool was validated and applied on large-scale smFISH image data sets, revealing diverse subcellular RNA localization patterns and a surprisingly high degree of cell-to-cell heterogeneity.

Hydrogenated microcrystalline silicon (µc-Si:H) and epitaxial silicon (epi-Si) films have been produced from SiF4, H2 and Ar mixtures by plasma enhanced chemical vapor deposition (PECVD) at 200 °C. Here, both films were produced using identical deposition conditions, to determine if the conditions for producing µc-Si with the largest crystalline fraction (XC), will also result in epi-Si films that encompass the best quality and largest crystalline silicon (c-Si) fraction. Both characteristics are of importance for the development of thin film transistors (TFTs), thin film solar cells and novel 3D devices since epi-Si films can be grown or etched in a selective manner. Therefore, we have distinguished that the H2/SiF4 ratio affects the XC of µc-Si, the c-Si fraction in epi-Si films, and the structure of the epi-Si/c-Si interface. Raman and UV-Vis ellipsometry were used to evaluate the crystalline volume fraction (Xc) and composition of the deposited layers, while the structure of the films were inspected by high resolution transmission electron microscopy (HRTEM). Notably, the conditions for producing µc-Si with the largest XC are different in comparison to the fabrication conditions of epi-Si films with the best quality and largest c-Si fraction.

Uveal melanoma (UM) is a rare tumor, yet it is the most common intraocular malignancy in adults with 500 new cases per year in France. It arises from the malignant transformation of melanocytes of the uvea, composed of the choroid (90% of cases), ciliary body (6%) and iris (4%). UM is a genetically simple tumor characterized by two major somatic driver events: mutually exclusive mutations of GNAQ, GNA11, PLCB4 or CYSLTR2, initiating tumorigenesis by constitutive activation of the Gaq signaling pathway, and a second event through mutually exclusive mutations that define metastatic risk and patient outcome: loss-of-function mutations of BAP1 associated with its bi-allelic inactivation by monosomy 3 (M3) leads to UM with high metastatic risk and poor prognosis, while disomy 3 (D3) is associated with either SF3B1 or EIF1AX mutations and respectively confer intermediate and low metastatic risk. Prognosis is dismal after occurrence of metastasis in 30-50% of cases, almost invariably to the liver, with a median survival of ∼12 months. Strikingly, individuals of European ancestry with light eyes and fair skin are particularly at risk of developing UM, with a relative risk (RR) of up to 20 in populations of European ancestry compared to those of African or Asian ancestries. However, the absence of ultra-violet (UV) mutational signature and stable UM incidence over the past decades rule out a role for pigmentation protecting against UV light to explain this peculiar epidemiology. Besides, familial cases of UM are seen in 1% of UMs, yet germline mutations of BAP1 are the only strongly predisposing genetic risk factor known so far and only explain a fraction of familial cases, suggesting that genetic risk of UM mostly remains unaccounted for.We hypothesized that some susceptibility alleles could predispose populations of European ancestry to UM. We sought to investigate the genetic predisposing factors in UM following two approaches: (i) a candidate-gene approach by targeted-sequencing of MBD4 in a large consecutive cohort of 1,093 UM patients, allowing us to demonstrate that MBD4 is a UM predisposing gene associated with high-risk M3 UM, conferring a RR = 9.2 of developing UM compared to general population; and (ii) a genome-wide association study in UM followed by a functional genomics study to identify common, low-penetrant UM risk alleles. Through this second approach, we confirmed the UM risk region at chromosome (chr) 5 on the TERT/CLPTM1L locus and further identified two pigmentation susceptibility loci, HERC2/OCA2 (ch15) and IRF4 (chr6). Importantly, these two loci were differentially associated with high-risk, M3 and low-risk, D3 UMs respectively, distinguishing two UM subgroups marked by specific genetic risk factors influencing tumor biology and metastatic potential. Lastly, we functionally characterized the TERT/CLPTM1L UM risk locus. After demonstrating an allele-specific regulation of gene expression at this locus, we narrowed down the genomic region to one functional variant, rs452384, which exhibited allele-specific nuclear factor binding properties. Using a quantitative proteomic approach, we evidenced binding of NKX2.4 transcription factor specifically to the rs452384-T allele. We further show that knockdown of NKX2.4 in UM cell lines results in a subtle yet significant increase in TERT and CLPTM1L expression, representing the first steps towards deciphering the tumorigenic mechanism of this risk locus in UM. Finally, we suggest that rs452384 may mediate its activity through differential telomere length regulation, potentially arising prior to UM tumorigenesis. Taken together, these results shed light on multiple moderate and low-risk genetic predisposition factors in UM, which bring new insights on the underlying biological mechanisms in UM predisposition, and has clinical relevance as patients with MBD4 deficiencies may respond to immunotherapy, resulting in the addition of MBD4 to clinical genetics panels., Le mélanome uvéal (MU) est un cancer rare, mais reste la tumeur intraoculaire la plus fréquente chez l’adulte avec 500 nouveaux cas par an en France. Le MU est lié à la transformation maligne des mélanocytes de l’uvée, composée de la choroïde (90% des cas), du corps ciliaire (6%) et de l’iris (4%). Le MU a une faible charge mutationnelle et comporte deux événements oncogéniques majeurs : l’activation constitutive de la voie de signalisation Gaq par mutations activatrices mutuellement exclusives de GNAQ, GNA11, PLCB4 ou CYSLTR2, et une seconde étape dans la transformation maligne, caractérisée elle aussi par des mutations mutuellement exclusives impactant le risque métastatique : les mutations de BAP1 (associées à une inactivation bi-allélique du gène par perte de l’allèle sauvage par monosomie 3, M3) constituent le MU à haut risque métastatique et mauvais pronostic, tandis que les mutations de SF3B1 ou EIF1AX (associées à la disomie 3, D3) confèrent un risque respectif intermédiaire et faible de métastases. Le très mauvais pronostic du MU métastatique, hépatique dans la majorité des cas, se caractérise par une survie médiane de 12 mois. Le MU est associé aux populations européennes, et aux individus à la peau et aux yeux clairs, avec un risque relatif (RR) 10 à 20 fois plus élevé chez les populations de souche européenne qu’africaine ou asiatique. L’absence de signature UV exclue le rôle des différences de pigmentation dans la protection des UV pour expliquer cette épidémiologie remarquable. De même, les mutations germinales de BAP1, seul facteur de risque génétique à forte pénétrance connu, n’expliquent qu’une partie des cas familiaux du MU, suggérant l’existence d’autres facteurs de prédisposition génétique.Nous avons émis l’hypothèse d’allèles à risque prédisposant les populations européennes au MU. Nous avons utilisé deux approches : (i) l’évaluation du gène candidat MBD4 par criblage de mutations dans une cohorte consécutive de 1,093 patients de MU, nous permettant de démontrer que MBD4 prédispose au MU à haut risque métastatique avec un RR de 9 comparé à la population générale ; et (ii) l’identification de polymorphismes fréquents de risque par étude d’association pan-génomique et la caractérisation de leurs conséquences moléculaires par une étude de génomique fonctionnelle. Nous avons ainsi pu démontrer l’implication de régions de risque sur les chromosomes 5, 6 et 14 (loci respectifs TERT/CLPTM1L, IRF4 et HERC2/OCA2). Les deux derniers loci ont un rôle majeur dans la pigmentation ; nous avons pu montrer leur association différentielle avec le MU à haut (M3) et faible (D3) risque métastatique (HERC2/OCA2 et IRF4 respectivement), indiquant deux facteurs de risque distincts ayant un impact sur le potentiel métastatique. Enfin, la caractérisation fonctionnelle du locus TERT/CLPTM1L nous a permis de démontrer la spécificité allélique de la régulation génique à cette région et d’identifier un variant fonctionnel du MU : rs452384. Celui-ci permet la fixation différentielle de facteurs nucléaires, parmi lesquels nous avons identifié par spectrométrie de masse quantitative NKX2.4, facteur de transcription spécifiquement associé à l’allèle T de rs452384. La sous-expression de NKX2.4 dans des modèles cellulaires du MU résulte en l’augmentation légère mais significative de l’expression des gènes TERT et CLPTM1L, constituant la première étape dans l’élucidation des mécanismes biologiques de cette région de risque. Enfin, nous émettons l’hypothèse que rs452384 régule de manière différentielle la longueur des télomères, avant l’apparition de la tumeur. Ces résultats mettent en avant le rôle de facteurs de risque à pénétrance faible ou modérée dans le MU et l’implication potentielle de mécanismes biologiques sous-jacents. Notamment, l’impact clinique du MU dans un contexte d’inactivation de MBD4 est l’inclusion de MBD4 dans les panels oncogénétiques pourraient permettre l’identification de patients répondant à l’immunothérapie.

International audience

Bovine cruor, a slaughterhouse waste, is produced in large quantities all around the world. This co-product was mainly composed of hemoglobin, a protein rich in bioactive peptides after its enzymatic hydrolysis. However, during conventional hydrolysis, chemical agents are necessary to adjust/regulate the pH of the solution and the final hydrolysates produced contain high levels of mineral salts. Therefore, in this study, it is proposed to apply, for the first time, a green technology, named electrodialysis with bipolar membrane (EDBM), as an alternative method to the conventional enzymatic hydrolysis of hemoglobin to obtain purified bioactive peptides. The main objectives of the present thesis were to test the feasibility of this new process to produce bioactive peptides from bovine hemoglobin, to establish the optimal conditions, to avoid membrane fouling and to apply a new original « multiple-step » EDMB process allowing the production of demineralized bioactive peptides without the addition of chemical salts. Bipolar/monopolar (anionic or cationic) configurations using the H+ and OH- generated by the bipolar membranes to regulate the pH were investigated and compared to a conventional process using chemical acid and base. The EDBM configuration formed with cationic membranes allowed the production of hydrolysates containing a low concentration of mineral salts but with fouling formation on the cationic membrane, while EDBM configuration formed with anionic membranes allowed the production of hydrolysates without fouling but with a similar salt concentration than the control. Based on these results, a new 3 compartments EDBM configuration was carried-out for denaturing the hemoglobin, inactivating the enzymatic reaction and demineralizing up to 85% the hemoglobin hydrolysate simultaneously. However, a fouling was still observed on the anionic membrane due to hem precipitation. For this reason, an additional step of discoloration was tested before the demineralization to avoid fouling using the electrogenerated acid. The discolored and demineralized peptides recovered showed antioxidant activity, antibacterial activity against many bacterial strains (Gram + and Gram -) and for the first time antifungal activity against many molds and yeasts strains. Moving towards a circular economy, this sustainable technology has found to be effective in performing multiple operations simultaneously and has a great potential for industrial hydrolysis of blood, since it produces purified biopeptides with a low mineral content and can be used as natural preservatives on meat.; Le cruor bovin, déchet des abattoirs est produit en très grande quantité dans le monde. Ce coproduit est composé principalement d'hémoglobine, une protéine riche en peptides bioactifs après son hydrolyse enzymatique. Cependant, lors de l'hydrolyse conventionnelle par la pepsine, des agents chimiques sont nécessaires pour ajuster et réguler le pH de la solution et, les hydrolysats finaux produits contiennent des niveaux élevés en sels minéraux. Pour pallier ces inconvénients, il a été proposé d'appliquer dans cette étude, pour la première fois, une technologie verte, appelée électrodialyse avec membrane bipolaire (EDMB), comme méthode alternative à l'hydrolyse enzymatique conventionnelle de l'hémoglobine afin d’obtenir des peptides bioactifs purifiés. Les objectifs de cette thèse étaient de tester la faisabilité de ce nouveau procédé pour la production de peptides bioactifs à partir d'hémoglobine bovine, d'établir les conditions optimales, d'éviter le colmatage membranaire et d'appliquer un nouveau procédé original d’EDMB à « multiple-étapes » permettant la production de peptides bioactifs déminéralisés sans ajout de produits chimiques. Des configurations bipolaires/monopolaires (anioniques ou cationiques) utilisant les ions H+ et OH- générés par les membranes bipolaires pour réguler le pH ont été étudiées et comparées à un procédé conventionnel utilisant des acides et des bases chimiques. La configuration d’EDMB formée avec les membranes cationiques a permis la production d'hydrolysats contenant une faible concentration en sels minéraux mais avec la présence d’un colmatage sur la membrane cationique, alors que la configuration d’EDMB utilisant des membranes anioniques a permis la production d'hydrolysats sans colmatage mais avec une concentration en sel similaire à celle de l’hydrolyse conventionnelle (contrôle). En se basant sur ces résultats, une nouvelle configuration d’EDMB à 3 compartiments a été mise en place et étudiée pour dénaturer l'hémoglobine, inactiver la réaction enzymatique et déminéraliser à 85% l'hydrolysat peptidique en simultané. Cependant, un colmatage a encore été observé sur la membrane anionique en raison de la précipitation de l'hème. Pour cette raison, une étape supplémentaire de décoloration a été réalisée avant la déminéralisation pour éviter le colmatage en utilisant l'acide électro-généré. Les peptides décolorés et déminéralisés récupérés ont montré une activité antioxydante, une activité antibactérienne contre plusieurs souches bactériennes (Gram + et Gram -) et pour la première fois une activité antifongique contre de nombreuses souches de moisissures et de levures. Dans un contexte d’économie circulaire, cette technologie durable s'avèrerait efficace pour effectuer plusieurs opérations simultanément et présente un potentiel important au niveau industriel pour l'hydrolyse du sang, puisqu'elle produit des bio-peptides purifiés ayant une faible teneur en sels minéraux et pouvant être utilisés comme conservateurs naturels sur la viande.

Extracellular vesicles (or EVs) are heterogeneous nano-sized cellular subspecies released by all cells in the body. They circulate freely in biological fluids and are today considered as potential markers and indicators of many diseases. Thus, identifying and characterizing them remains a major challenge that could lead to rapid diagnosis and prognosis of many diseases. However, due to their heterogeneous sizes, no current technique allows for their proper quantification and characterization. Thus, only a combination of technologies could provide important information on these molecules of interest. To this end, this thesis aims at developing a lab on chip that would allow the isolation of EVs from a biological fluid. To achieve this separation, we chose to use hydrodynamic filtration. This principle allows the separation of particles according to a defined size. It is a simple and passive method (no external field), robust and less prone to clogging than standard filtration techniques. Its principle is based on the size exclusion of particles whose diameter is greater than a defined value, called the cut-off radius, monitored by the ratio of flow rates between a main channel and one or more side channels. We have designed and manufactured microfluidic chips allowing the separation of particles at a cut-off radius adapted to the EV application (between 150 and 900 nm). We used an original microfabrication technique based on successive dry film lamination (DF) and photolithography to manufacture the devices. We demonstrated the proof of concept using three fluorescent biofunctionalized nanoparticles (NP140, N480 and NP 920) covering the EV distribution range. Once the validation was done in synthetic media, we injected biological media spiked with nanoparticles into the sorting devices. We were able to quantify and analyze filtered samples using different techniques (Tunable Resistive Pulse Sensing, flow cytometry, direct counting). We have demonstrated that using our device we were able to collect 35 µL of filtered sample within 30 minutes which is highly compatible with the requirements of the subsequent quantification analysis. We compared filtered samples to injected samples and observed identical concentrations for the nanoparticles (NP 140-CAS, NP 480-OVA) smaller than the cut-off radius, indicating a 100% recovery yield. For nanoparticles larger than the cut off radius (NP 920) and cells present in the biological media (red and white blood cells, etc.) we obtained a filtration yield of the order of 96 to 98 %. We thus demonstrate that the use of biological fluid (blood, plasma, PPL), does not affect the separation results in any way. These results proved that this separation method is a good candidate for treatment of natural samples and could serve as a new range of diagnostic and prognostic tests for pathologies., Les vésicules extracellulaires (ou EVs) sont des sous-espèces cellulaires, de tailles nanométriques et hétérogènes, libérées par toutes les cellules de l'organisme. Elles circulent librement dans les fluides biologiques et sont aujourd'hui considérées comme de potentiels marqueurs et indicateurs de nombreuses maladies. Ainsi, les identifier et les caractériser reste un défi majeur qui pourrait conduire à des diagnostics et pronostics rapides de nombreuses pathologies. Cependant, de par leurs tailles hétérogènes, aucune technique actuelle ne permet de bien les quantifier et les caractériser. Ainsi, seule une combinaison de technologies pourrait permettre de fournir des informations importantes sur ces espèces d'intérêts. Dans ce but, ce travail de thèse a pour objectif de développer un laboratoire sur puce qui permettrait d'isoler les EVs à partir d'un fluide biologique. Le laboratoire sur puce est basé sur un principe de filtration hydrodynamique permettant la séparation des EVs provenant de milieux complexes. Ce principe permet de séparer les particules selon une taille définie. Il s'agit d'une méthode simple et passive (aucun champ extérieur), robuste et moins sujette au risque de colmatage. Son principe est basé sur l'exclusion en taille des particules dont le diamètre est supérieur à une valeur définie, appelée le rayon de coupure, définie par le rapport des débits entre une canalisation principale et une ou plusieurs canalisations latérales. En choisissant judicieusement les dimensions des canalisations, il a été possible de concevoir plusieurs dispositifs permettant la séparation à un rayon de coupure donné. Nous avons conçu et fabriqué, des puces microfluidiques permettant la séparation des particules à un rayon de coupure adapté à l'application des EVs (entre 150 et 900 nm). Nous avons utilisé une technique originale basée sur le laminage successif de films secs (DF) et la photolithographie. Nous avons fait la preuve de concept à l'aide de trois nanoparticules biofonctionnalisées fluorescentes (NP 140-CAS, N 480-OVA et NP 920) couvrant la gamme de distribution des EVs. Une fois la validation faite dans des milieux synthétiques, nous avons décidé de complexifier les mélanges et nous avons injecté des milieux biologiques associés à de nanoparticules dans les dispositifs de tri. Une fois les différentes séparations effectuées, nous avons pu quantifier et analyser à l'aide de différentes techniques (par Tunable Resistive Pulse Sensing, par cytométrie de flux, par comptage directe) nos échantillons. Nous avons démontré qu'à l'aide de nos dispositifs, nous pouvons récolter 35 µL d'échantillon filtré en 30 minutes d'injection ce qui est hautement compatible avec les exigences de l'analyse de quantification. Nous avons comparé chaque échantillon filtré avec ceux injectés et nous avons observé des concentrations identiques pour les nanoparticules de taille inférieure (NP 140-CAS, NP 480-OVA) au rayon de coupure, nous indiquant un rendement de 100 % de récupération. Pour les autres nanoparticules (NP 920) et les grosses cellules présentes dans les milieux biologiques (globules rouges, blancs, etc...) nous avons obtenu un rendement de filtration de l'ordre de 96 à 98%. Nous démontrons ainsi que l'utilisation de fluide biologique (sang, plasma, PPL), n'affecte en aucun cas les résultats de séparation. Ceci prouve que cette méthode de séparation est un bon candidat pour le traitement des échantillons naturels et pourrait être la base d'une nouvelle génération de tests diagnostiques.

Escherichia coli is a widespread cellular host for the industrial production of protein-based biopharmaceuticals. This production, mainly operated in fed-batch mode, aims to maximize biomass productivity. However, the accumulation of acetate during the culture inhibits the cells respiratory capacity and lowers their metabolic performance. In this thesis, closed-loop feeding control strategies are considered to avoid acetate accumulation and maximize biomass productivity. To this end, modelbased control and estimation schemes are developed to regulate the biomass growth rate and the acetate concentration. The control methods ranged from the Generic Model Control and Nonlinear Model Predictive Control, and the non-measured variables are estimated using the Unscented Kalman Filter. The developments focused on the robustness of the proposed methods due to the uncertain nature of the bioprocess. The performance and robustness of the control and estimation strategies are tested and tuned by means of different scenarios of simulation runs. Fed-batch cultures of E. coli BL21(DE3) strain are successfully carried on a lab-scale bioreactor, highlighting the potential of the proposed strategies in real-time conditions. The proposed control strategies presented in this thesis lead to an average gain of up to 20% in biomass productivity compared to the conventional operating mode., Escherichia coli est un hôte cellulaire très répandu pour la production industrielle de produits bio-pharmaceutiques à base de protéines. Cette production, principalement opérée en mode fed-batch, vise à maximiser la productivité de la biomasse. Cependant, l'accumulation d'acétate durant la culture inhibe la capacité respiratoire des cellules et diminue leur performance métabolique. Dans cette thèse, des stratégies de commande de l'alimentation sont envisagées pour éviter l'accumulation d'acétate et maximiser la productivité de la biomasse. A cette fin, des schémas de commande et d'estimation à base de modèle sont développés pour réguler le taux de croissance de la biomasse et la concentration en acétate. Les méthodes de commande vont du contrôle par modèle générique au contrôle prédictif par modèle non linéaire, et les variables non mesurées sont estimées à l'aide du filtre de Kalman "sans parfum". Les développements ont porté sur la robustesse des méthodes proposées en raison de la nature incertaine du bioprocédé. La performance et la robustesse des schémas de commande et d'estimation sont testées et ajustées au travers de différents scénarios de simulation. Des cultures en mode Fed-batch de la souche E. coli BL21(DE3) sont réalisées avec succès sur un bioréacteur de laboratoire, mettant en évidence le potentiel des stratégies proposées dans un contexte de conditions opératoires en temps réel. Les stratégies de commande proposées dans cette thèse permettent un gain moyen jusqu'à 20% de la productivité de la biomasse par rapport au mode de fonctionnement conventionnel.

International audience; Light is a key environmental factor for the growth of micro-algae, and optimizing the capture of light is critical for high efficiency production systems. As the density of the population of micro-algae increases, the availability of light decreases, leading to a reduction in the growth rate because of mutual shading, while other effects, such as photo-inhibition, might be especially frequent when the population density is low. Several models in the literature have been developed to take into account light phenomena and predict micro-algal growth, particularly in a mono-culture. With the help of a simple expression for the attenuation of the light, we propose and justify a new growth function that incorporates both photo-inhibition and photolimitation. In agreement with the experimental data, this new formulation describes the micro-algal response to a wide range of situations of excessive or insufficient light intensities through an explicit dependence on both the incident light and the biomass concentration. While simple, the proposed expression can be satisfactorily applied to practical cases under nutrient replete conditions in photo-bioreactors with different sizes and geometries. It extends naturally to the growth of different species, providing a dynamic model which can simulate experiments in a mono-culture as well as in polycultures. The investigation of the competition for light-limited growth shows that the model predicts competitive exclusion, which has also been experimentally demonstrated. This leads to new perspectives for the control and optimization of mixed micro-algal cultures.

An in situ Modulated Photoluminescence (MPL) measurement setup mounted on a Plasma Enhanced Chemical Vapor Deposition (PECVD) reactor is described. A method for deriving the actual minority carrier lifetime at the specific injection level of 1.1015 cm−3 is presented. This tool was used in a case study to monitor the passivation properties of aluminum oxide (AlOx) thin films upon annealing. Interesting kinetics, such as a drop of lifetime at temperatures higher than 250 °C along with a recovery of the lifetime during the cooling phase, are shown. Moreover, these in situ results combined with ex situ studies allowed us to demonstrate the formation a stable high positive fixed charge density in the AlOx layer (+1.1012 cm−2) as a consequence of the combination of annealing and light exposure.

International audience; A Gram-positive, anaerobic coccus isolated from a human surgical site infection was previously shown to belong to an unknown species of the genus Peptoniphilus initially proposed as ‘Peptoniphilus nemausus’ sp. nov., based on both 16S rRNA gene sequence identity of 97.9% with the most closely related species Peptoniphilus coxii and an individualized phylogenetic branching within the genus Peptoniphilus. A polyphasic characterization of the novel species is proposed herein. Whole genome sequence analysis showed an average nucleotide identity value of 84.75% and digital DNA-DNA hybridization value of 28.9% against P. coxii type strain. The strain displayed unique features among members of the genus Peptoniphilus, as it was able to hydrolyze aesculin, and produced acetate as the major metabolic end-product without associated production of butyrate. Growth was observed under microaerophilic conditions. From all these data, the isolate is confirmed as belonging to a new Peptoniphilus species, for which the name Peptoniphilus nemausensis sp. nov. is proposed. The type strain is 1804121828T (= LMG 31466T = CECT 9935T). A database survey using a highly polymorphic partial sequence of the 16S rRNA gene of P. nemausensis revealed P. nemausensis to be a particularly rare skin-associated species in humans. An emendated description of the Peptoniphilus genus is proposed based on a review of the characteristics of the 12 new species with validly published names since the genus description in 2001 and of P. nemausensis. Finally, the relationships between members of the genus Peptoniphilus were explored based on whole genome sequence analysis in order to clarify the taxonomic status of not yet validly published species showing that three pairs of species should be considered as synonyms: Peptoniphilus timonensis and ‘Peptoniphilus phoceensis’, Peptoniphilus lacydonensis and ‘Peptoniphilus rhinitidis’, Peptoniphilus tyrrelliae and Peptoniphilus senegalensis.