IRT Jules Verne

1. RasGRP1 Regulation in Autoimmunity T cell selection is essential to allow maturation only of capable and not auto-reactive T cells. Our research demonstrates that reduced expression of RasGRP1 (a key factor in T cell receptor signaling) leads to defects in T cell selection and autoimmunity in mice. Reduced expression of RasGRP1 was observed in T cells of patients with rheumatoid arthritis, with a more pronounced reduction in active disease. We identified a regulatory mechanism of RasGRP1 expression where the transcription factor RUNX1 interacts with an enhancer region containing SNPs associated with autoimmunity. RUNX1 was found to be less expressed in T cells of patients with rheumatoid arthritis as well, suggesting involvement in RasGRP1 regulation and autoimmunity. 2. Improved Single-Cell Segmentation for Imaging Mass Cytometry data We developed a novel method called MATISSE to improve the segmentation quality of single cells in tissue images captured using a multiplex technology called Imaging Mass Cytometry (IMC). We apply a fluorescence microscopy step preceding IMC imaging, boosting the identification of individual cell nuclei at higher resolution. This vastly reduces cell segmentation artifacts such as merged and fragmented events. With this approach, more complete and reliable single-cell datasets can be generated, which is essential to study tissue biology. 3. Inflammatory Bowel Disease & Colitis-Associated Cancer Long-standing active inflammatory bowel disease (IBD) increases risk of developing colorectal cancer, specifically colitis associated cancer (CAC). We apply several technologies, including IMC to characterize the immune context of CAC development in uninflamed, inflamed, and early dysplastic samples of IBD patients. Key findings include a robust increased type 3 inflammation (IL-17) and JAK-STAT signaling (P-STAT3) in inflamed and early dysplastic tissue, while cytotoxic T cell surveillance of epithelial cells appears reduced compared to the uninflamed condition. 4. Colorectal Cancer Metastasis We investigated the tumor microenvironment in high microsatellite instability (MSI-H) colorectal primary tumors in the context of metastatic capacity. By means of multiplex tissue imaging we revealed distinct tissue composition in metastatic versus non-metastatic tumors. In particular primary tumors with metastatic potential display a high stromal content and low influx of cytotoxic T cells, which may serve as predictive markers for metastasis formation.

Abstract. Wind blades are usually fabricated with the vacuum-assisted resin infusion process (VARTM), using thermoset resins, which exhibit recycling limitations. The ZEBRA project, led by the IRT Jules Verne, aims to accelerate circular economy by fabricating recyclable wind blades, regrouping major industrial players: LM Wind Power, Arkema, CANOE, Owens Corning, ENGIE and Suez (Fig. 1). Therefore, thermosets are replaced with the recyclable thermoplastic resin Elium® developed by ARKEMA. The fabrication of thick, large and complex components using reactive resins presents various physical and geometrical difficulties, such as infusion strategy, and the polymerization behavior of the Elium® resin. This paper addresses the use of simulation to predict the infusion and polymerization process for parts of different complexities.

An innovative simultaneous process, using supercritical fluid (SCF) chemistry, was used to recycle uncured prepregs and to functionalize the recovered carbon fibres with Fe3O4 magnetic nanoparticles (MNPs), to produce a new type of secondary raw material suitable for composite applications. This specific functionalization allows the fibres to be heated by induction through a hysteresis loss mechanism characteristic for nanoparticle susceptor-embedded systems, for triggered healing properties and a potentially easy route for CF reclamation. Using SCF and hydrothermal conditions for recycling, functionalization of fibres can be performed in the same reactor, resulting in the creation of ready-to-use fibres and limiting the use organic solvent. After cutting the uncured prepreg to the desired length to fit in future applications, supercritical CO2 extraction is performed to partially remove some components of the uncured prepreg matrix (step 1). Then, the recycled carbon fibres (rCFs), still embedded inside the remaining organic matrix, are brought into contact with reactants for the functionalization step (step 2). Two possibilities were studied: the direct synthesis of MNPs coated with PAA in hydrothermal conditions, and the deposition of already synthesized MNPs assisted by supercritical CO2-acetone. No CF surface activation is needed thanks to the presence of functional groups due to the remaining matrix. After functionalization, ready-to-use material with homogeneous depositions of MNPs at the surface of rCF is produced, with a strong magnetic behaviour and without observed degradation of the fibres.

Abstract This papers derives finite sample results to assess the consistency of Generalized Pareto regression trees introduced by Farkas et al. [2021] as tools to perform extreme value regression for heavy-tailed distributions. This procedure allows one to constitute clusters of similar tail behaviors depending on the value of the covariates, based on a recursive partition of the sample and simple model selection rules. The results we provide are obtained from concentration inequalities, and are valid for a finite sample size. A misspecification bias that arises from the use of a Peaks over Threshold approach is also taken into account. Moreover, the derived properties legitimate the pruning strategies, that is the model selection rules, used to select a proper tree that achieves a compromise between simplicity and good fit. The methodology is illustrated through a simulation study, and a real data application in insurance against natural disasters.