Archives ouvertes de Paris-Saclay
Not a member yet
272730 research outputs found
Sort by
Parasitic connections: a patescibacterial epibiont, its methylotrophic gammaproteobacterial host, and their phages
No full textInternational audienceABSTRACT Patescibacteriota form a very diverse and widely distributed phylum of small bacteria inferred to have an episymbiotic lifestyle. However, the prevalence of this lifestyle within the phylum and its host specificity remain poorly known due to the scarcity of cultured representatives. Here, we describe a complex system consisting of a patescibacterium, its gammaproteobacterial hosts, and their respective phages based on enrichment cultures and metagenomic data from two shallow, geographically close, freshwater ecosystems. The patescibacterium Strigamonas methylophilicida sp. nov. defines a new genus within the family Absconditicoccaceae. It grows as an epibiont on cells of methanotrophic species of the gammaproteobacterial family Methylophilaceae. Strigamonas cells grow tightly attached to the host, sometimes forming stacks that connect two host cells. Despite a surprisingly large genome (1.9 Mb) compared to many other Patescibacteriota, S. methylophilicida lacks many essential biosynthetic pathways, including the complete biosynthesis of phospholipids, amino acids, and nucleic acids, implying a dependence on the host to obtain these molecules. We also identified and assembled the complete genomes of one patescibacterial phage that might represent a new virus family within the class Caudoviricetes , and two Methylophilaceae phages predicted to have head-tailed and filamentous virions, respectively. The patesciphage uses a modified genetic code similar to that of its host and encodes four tRNA genes, including the suppressor tRNA gene for the UGA stop codon, which is reassigned to glycine in many Patescibacteriota. Our results confirm a prevalent episymbiotic lifestyle in Absconditicoccaceae and further suggest a clade-specific adaptation of this patescibacterial family for gammaproteobacterial hosts. IMPORTANCE Patescibacteriota are ultra-small bacteria with reduced genomes that rely on symbiotic interactions with other prokaryotes; however, their host specificity and associated viral parasites remain poorly characterized due to limited cultured representatives. By combining targeted cultivation with genomic and microscopy analyses, we reveal previously unrecognized host lineages and expand the known viral diversity infecting this major, but still poorly known, bacterial phylum. We describe Strigamonas methylophilicida , a new patescibacterial species of the family Absconditicoccaceae that grows as an epibiont on various methylotrophic Gammaproteobacteria. This expands the host range for this family, previously found to infect only photosynthetic partners. Using enrichment cultures and metagenomics, we retrieved complete genomes of novel phages infecting S. methylophilicida and its methylotrophic hosts, including one phage that uses a modified genetic code matching that of the patescibacterium, which shows a specific viral adaptation to infect Absconditicoccaceae hosts. Our findings reveal a previously unrecognized patescibacteria-methylotrophs-phages tripartite interaction in freshwater environments, highlight the adaptations of patescibacterial phages, and shed light on the complex ecology and evolution of host-parasite-phage dynamics in understudied bacterial lineages
Apprentissage automatique et analyse de données à hautes performances pour la conception de chemins de fer de nouvelle génération
No full textThe MINERVE project aims to create a digital twin of the French railway infrastructure,which will then be used to assess the state of the railway and predict its evolution. To achieve such goal, the project requires a robust architecture capable of collecting, storing, and analyzing large amounts of data from various sources. During my thesis, we developed a theoretical big data architecture tailored to MINERVE's objectives, based on the Zaloni zone data lake architecture, designed to be implemented on the Paris-Saclay mesocenter, a high-performance computing facility.We then used this theoretical architecture to implement two use-cases that would support the development of MINERVE Digital Twins. The first use case is the development of a pipeline for the semantic segmentation of large railway point clouds, a needed work to enhance the automationof railway infrastructure monitoring and maintenance. The second use case focuses on the generation of high-frequency seismic waves from low-frequency ones, a necessary step in the making of an accurate Digital Twin, as the ability to detect earthquakes and other seismic events is crucial for the safety and reliability of railway operations.Le projet MINERVE vise à créer un jumeau numérique de l'infrastructure ferroviaire française,qui permettra ensuite d'évaluer l'état du réseau et de prédire son évolution. Pour atteindre cet objectif, le projet nécessite une architecture robuste capable de collecter, stocker et analyser de grandes quantités de données provenant de sources diverses.Au cours de ma thèse, nous avons développé une architecture big data théorique adaptée aux objectifs de MINERVE, basée sur l'architecture du data lake en zone Zaloni, conçue pour être implémentée sur le mésocentre Paris-Saclay, une plateforme de calcul haute performance.Nous avons ensuite utilisé cette architecture théorique pour implémenter deux cas d'utilisation qui soutiendraient le développement du jumeau numérique de MINERVE.Le premier cas d'utilisation est le développement d'un pipeline pour la segmentation sémantique de grands nuages de points ferroviaires, un travail nécessaire pour améliorer l'automatisation de la surveillance et de la maintenance de l'infrastructure ferroviaire. Le deuxième cas d'utilisation porte sur la génération d'ondes sismiques haute fréquence à partir d'ondes basse fréquence, une étape nécessaire à la création d'un jumeau numérique précis, car la capacité à détecter les tremblements de terre et autres événements sismiques est cruciale pour la sécurité et la fiabilité des opérations ferroviaires
Extended Fused Carbazole‐BODIPY, High Brightness NIR Organic Dyes
No full textInternational audienceFluorescence‐based bioimaging enables noninvasive visualization of molecular and cellular processes with high sensitivity and without ionizing radiation. However, conventional fluorophores emitting in the visible or near‐red infrared I (NIR‐I) (650–800 nm) regions suffer from limited tissue penetration and scattering. Extending fluorescence emission into the deeper NIR region represents a promising strategy to overcome these drawbacks, yet achieving high brightness and stability in organic dyes remains a major challenge. We report an original family of hetero‐substituted‐fused boron‐dipyrromethene (BODIPY) dyes bearing carbazole and thienyl donors that exhibit record brightness and emission maxima up to 852 nm in toluene. The synthetic route combines successive Stille couplings from a 2,6‐dibromo‐3,5‐diiodo‐BODIPY precursor and an unprecedented silver(I)‐mediated oxidative cyclization, affording high yields and suppressing undesired chlorination. The resulting dyes display intense absorption ( ε = 1.8–2.5 × 10 5 M − 1 cm − 1 ) and exceptional fluorescence quantum yields ( Φ up to 0.73). Encapsulation in silica nanoparticles (NPs) preserves their photophysical properties and enables efficient NIR‐II in vivo imaging in mice, allowing tumor detection at doses as low as 0.2 nmol with tumor‐to‐muscle ratios > 4. These fused BODIPY derivatives rank among the brightest NIR fluorophores reported to date and open new avenues for high‐contrast deep‐tissue imaging and image‐guided surgery
Impact of hydrogen implantation on a transferred diamond layer
No full textInternational audienceHydrogen implantation combined with bonding facilitates the transfer of thin films. This process, known as Smart Cut™, is well-established for silicon in the fabrication of SOI stacks but remains exploratory for diamond due to its rough and non-planar surface for bonding. To improve bonding energy, we used surface activation bonding in this article. This approach demonstrates a successful transfer of thin diamond films onto silicon with a yield of nearly 90%. However, the post-fracture diamond film exhibited a pyramidal surface topology. This particular topology is characterized by Raman spectroscopy, Cathodoluminescence, Scanning Electron Microscopy, Transmission Electron Microscopy, Atomic Force Microscopy, and Laue microdiffraction, and results from the formation of dihydrogen blisters. These blisters cause the formation of vertical graphite sheets on the film's surface and induce plastic deformation in the underlying silicon substrate without compromising the diamond film's crystallinity or the bonding. Additionally, we propose a post-fracture surface cleaning method to obtain an epi-ready film and enable the reuse of the donor substrate
Neutron Crystallography Study of Host–Pathogen Recognition Enhanced by Hydrogen/Deuterium Exchange on Carbohydrates
No full textInternational audienceNeutron macromolecular crystallography (NMX) is a unique tool for location of hydrogen atoms. Its application is greatly enhanced by the use of deuterium-labeled molecules. Human glycans are the targets of virulence factors from pathogens, such as the soluble lectins from the opportunistic bacterium Pseudomonas aeruginosa. Deuterated galactose was obtained by hydrogen isotope exchange and cocrystallized with a fully deuterated bacterial receptor, LecA, a lectin involved in P. aeruginosa tissue adhesion and biofilm formation. The structure of the complex determined using neutron diffraction reveals the positions of all hydrogen atoms, as deuterium, highlighting the important role of a charged histidine in the binding site, the bridging by a buried water molecule and the influence of the coordinating calcium ion on the adjacent hydrogen bonds. LecA is a target for pathoblockers and these structural details can assist in the design of glycomimetics for fighting multidrug resistant infections
Linking species traits and vulnerability indicators in European Odonata
No full textInternational audienceUnderstanding the mechanisms and commonalities driving species’ vulnerability is essential for prioritizing and guiding conservation efforts. Trait-based approaches offer a mechanistic foundation for generalizing species vulnerabilities within a taxonomic group. Here, we assess how the vulnerability of European Odonata is associated with their traits. Our aim was to (1) quantify the link between traits and vulnerability and (2) identify the most important traits in a multi-trait context. For 123 species, we linked 3 vulnerability indicators (Red List categories, distribution trends and areas of occupancy) to a dozen traits, using discriminant and redundancy analyses. We find that 48 to 64% of the variability in vulnerability indicators is explained by traits. The main traits related to vulnerability are habitat, voltinism and thermal preferences. More specifically, vulnerable species tend to associate with oligotrophic habitats or Mediterranean streams. They also tend to have longer life cycles, but this relationship is reversed for species with a small area of occupancy. Species vulnerable because of their decreasing distribution tend to have cold thermal preferences. Vulnerable species generally show a narrow thermal range (except for species vulnerable because of their decreasing distribution). Assessing species’ vulnerability is crucial to inform conservation: our trait-based approach provides clues regarding pressures responsible for species vulnerability, thus allowing to plan conservation action targeting groups of species sensitive to the same pressures, rather than focusing on individual species. Our method provides novel opportunities for predicting species’ vulnerability, and paves the way for building a multi-species conservation indicator for Odonata
Multidisciplinary science funding is more than ever a planetary priority: Reflections from the Make Our Planet Great Again (MOPGA) program
No full textInternational audienceGlobal change poses “wicked problems” that have become ever more complex, pervasive, and damaging. Developing innovative solutions increasingly require diverse research approaches. The Franco-German Make Our Planet Great Again (MOPGA) program was designed to create a unique international network of top-level research, from fundamental to solution-oriented projects. MOPGA stands out from other large research initiatives by focusing not on a singular central research challenge but on facilitating multidisciplinary interactions between traditionally separated fields. MOPGA recognized that social, natural and engineering sciences share a unifying aim to address global change. In addition to addressing timely and innovative research questions within disciplines, MOPGA worked to improve communication across disciplines via annual meetings for all laureates and their research groups, scientific board exchanges, and public online seminars. Drawing on our MOPGA experiences, we discuss how such exchanges should be extended to meet the needs identified by the scientific community, international policy-makers, and regional stakeholders. In the current political landscape of scientific suppression and heightened mistrust in scientific expertise, the need for such bold, independent and collaborative scientific initiatives is greater than ever
Adversarial Domain Adaptation Enables Knowledge Transfer Across Heterogeneous RNA-Seq Datasets
No full textAccurate phenotype prediction from RNA sequencing (RNA-seq) data is essential for diagnosis, biomarker discovery, and personalized medicine. Deep learning models have demonstrated strong potential to outperform classical machine learning approaches, but their performance relies on large, well-annotated datasets. In transcriptomics, such datasets are frequently limited, leading to over-fitting and poor generalization. Knowledge transfer from larger, more general datasets can alleviate this issue. However, transferring information across RNA-seq datasets remains challenging due to heterogeneous preprocessing pipelines and differences in target phenotypes. In this study, we propose a deep learning-based domain adaptation framework that enables effective knowledge transfer from a large general dataset to a smaller one for cancer type classification. The method learns a domain-invariant latent space by jointly optimizing classification and domain alignment objectives. To ensure stable training and robustness in data-scarce scenarios, the framework is trained with an adversarial approach with appropriate regularization. Both supervised and unsupervised approach variants are explored, leveraging labeled or unlabeled target samples. The framework is evaluated on three large-scale transcriptomic datasets (TCGA, ARCHS4, GTEx) to assess its ability to transfer knowledge across cohorts. Experimental results demonstrate consistent improvements in cancer and tissue type classification accuracy compared to non-adaptive baselines, particularly in low-data scenarios. Overall, this work highlights domain adaptation as a powerful strategy for data-efficient knowledge transfer in transcriptomics, enabling robust phenotype prediction under constrained data conditions
Bumblebees are the most efficient pollinators of raspberry and strawberry in urban environments
No full textInternational audiencePollinators are essential for crop pollination, but pollinators differ in their pollination efficiency. In urban areas, environmental filters such as soil sealing or the urban heat island lead to biotic homogenisation of pollinator communities, with generalist species being favoured while specialist species are filtered out. Therefore, efficient pollinators may be excluded from urban areas. In the context of the development of urban agriculture, urban areas require pollination by efficient pollinators. Here, we ask whether pollinators sustained in urban environments are equally efficient, and whether urbanisation impacts the efficiency of pollinators delivering pollination services.Using strawberry and raspberry as experimental plants, we carried out single visit experiments over 2 years (2023 and 2024) in spring and autumn, to assess pollination efficiency in densely urbanised and suburban sites in the region of Paris (France). We measured fruit mass, malformation and seed set of fruits which developed from flowers having received a single visit from a pollinator.3. We found that bumblebees were more efficient than honeybees as pollinators of raspberry, but not strawberry, as measured by fruit mass.Bumblebees were also more efficient than small and large solitary bees for pollinating strawberry, and more efficient than large solitary bees for pollinating raspberry. These differences were detected on mass for strawberries, and on mass, seed set and fruit malformation for raspberries.Practical implication: Environmental filters in urban environments tend to favour few generalist pollinator species. On one hand, these environments support honeybees, which we found were not necessarily the most efficient. On the other hand, wild pollinators, in particular bumblebees, were more efficient than honeybees for pollination of strawberry and raspberry. Thus, urban conservation strategies should focus on promoting these wild and efficient pollinators by planting beneficial plant species in flower beds and providing nesting habitats for ground nesting pollinators. This would promote diverse and efficient pollinators and thus enhance pollination services for urban agriculture