HAL Evry
Not a member yet
19237 research outputs found
Sort by
Comparison of sedimentary ancient DNA (sedaDNA) extraction and shotgun metagenomic library preparation techniques
No full textInternational audienceSedimentary ancient DNA (sedaDNA) is an emerging field, increasingly being applied to the study of past aquatic ecosystems. However, several sedaDNA extraction methods from deep-ocean sediment matrices and sequencing library preparation have recently been developed, which may complexify results comparison and interpretations. We present a sedaDNA interlaboratory comparison study to assess the impact of extraction and library preparation protocols on metagenomic results. We applied three extraction protocols to four samples from two sediment cores from the Northern Antarctic Peninsula: (1) a ‘combined’ protocol using ethylenediaminetetraacetic acid (EDTA) and silica-in-solution to isolate DNA, (2) a high-guanidine protocol involving long cold centrifugation to remove polymerase chain reaction (PCR) inhibitors, and (3) a commercial kit, the DNeasy PowerSoil Pro Kit. We also compared two library preparation protocols, both optimised versions from Meyer and Kircher (2010). Using a blind approach relying on k-mer similarity assessment, greater variability was observed between individual samples than between protocols. An in-depth analysis of eukaryotic and (highly abundant) diatom community composition revealed that all protocols recovered eukaryotic sedaDNA, with minor differences between extraction and library protocols on sequence composition. However, the different DNA extraction and library preparations influenced the DNA read length, potentially resulting in selective targeting of organisms with variable sedaDNA preservation. This study highlights the importance of selection and standardisation of protocols to ensure reproducibility and comparability of past ecosystem reconstructions, particularly at lower taxonomic levels, e.g. diatoms. Although complete standardisation across research projects is challenging, this study shows that data remain reasonably comparable when processed consistently
Orlanne, digital workeuse en mode "Feelgood"
No full textDeuxième partie. Émotionalité au travail : nouvelles préoccupations, réalités mouvantes et questionnements existentielsInternational audienc
Plants, Pollinators and Pheromones: Promises and Lies of Semiochemicals
No full textInternational audiencePollination is traditionally regarded as a quintessential mutualism, yet many plants employ deceptive strategies to achieve reproductive success. Among the most intriguing is sexual deception, wherein flowers mimic the sex pheromones and visual signals of female insects to attract male pollinators—without providing any reward. This strategy, most notably observed in orchids, is a powerful driver of diversification and speciation. Recent advances in genomics, metabolomics, and high‐resolution imaging are shedding light on the genetic and biochemical mechanisms underpinning these complex mimicry systems. Remarkably, subtle genetic modifications and the repurposing of existing gene networks can give rise to highly specialized and effective forms of deception. Central to this process are volatile organic compounds (VOCs), which serve as species‐specific semiochemicals that manipulate innate pollinator behaviors and reinforce reproductive isolation. This review synthesizes emerging insights into floral semiochemistry and highlights its broader applications in pollinator surveillance, crop pollination enhancement, and biodiversity monitoring. As global pollinator populations face increasing threats, understanding floral chemical ecology offers promising avenues for designing pollinator‐friendly crops and advancing tools in synthetic ecology
Seasonal Transition in the Dominance of Photoautotrophic and Heterotrophic Protists in the Photic Layer of a Subtropical Marine Ecosystem
No full textInternational audienceProtists are major functional players in the oceans. Time‐resolved protist diversity and succession patterns remain poorly described in subtropical ecosystems, limiting current understanding of food web dynamics and responses to environmental changes in these major world‐ocean regions. We used amplicon sequencing data and trait‐based annotation to examine the seasonality of planktonic protists in the subtropical Gulf of Aqaba (Red Sea). Temperature and nutrients were the major drivers of succession. We detected marked seasonal shifts in protists. Heterotrophs, including diverse parasitic functional groups, dominated the warm, stratified oligotrophic period spanning spring and summer. By contrast, nutrient influx during deep convective mixing in winter triggered a shift to photoautotrophic communities dominated by a few genera of chlorophytes. Deeper winter mixing resulted in larger blooms at the onset of stratification dominated by diatoms, relative to chlorophytes that prevailed during shallower blooms. This result illustrates the impact of mixing depth on bloom formation and composition. Comparisons with oceanwide rDNA datasets indicate that the oligotrophic protist assemblages from the Gulf resemble those from warm, open oceans. This work provides a detailed assessment of the seasonal switch in dominant trophic functions in protists in phase with nutrient levels in a subtropical planktonic ecosystem
"Gérer et militer", concevoir l'organisation militante avec François Rousseau
No full textCe travail en cours donne à voir les apports de la pensée de François Rousseau pour apprécier comment les organisations d'économie sociale et solidaire peuvent articuler deux actions au coeur de leurs modèles de gestion : gérer et militer. Dans ce texte, vous découvrirez le parcours atypique de Rousseau, la richesse de sa pensée pour apprécier l'importance de l'articulation entre un mythe, une communauté de militants et de militantes et les gestes sociaux qu'iels mettent en place pour assurer la pérennité d'une organisation d'économie sociale et solidaire. Sa pensée est pourtant méconnue. Pour autant, aujourd'hui, des auteurices mobilisent sa lecture anatlytique pour mieux comprendre ce qui se fabrique dans les organisations d'économie sociale et solidaire, comme les coopératives d'activités et d'emploi
La transformation numérique de la justice : ambitions, réalités et perspectives
No full textThe study, conducted over a four-year academic cycle with the assistance of M2 students from the Cyberjustice Master's programme at the Faculty of Law, Political Science and Management at the University of Strasbourg, aims to objectively assess the discourse and representations of the digital transformation of justice, in particular by capitalising on testimonials from professionals in the field and drawing on the available literature.L'étude, réalisée sur un cycle de quatre années universitaires avec le concours des étudiants du M2 du Master Cyberjustice de la Faculté de droit, de sciences politiques et de gestion de l’Université de Strasbourg, vise à objectiver les discours et les représentations de la transformation numérique de la justice, notamment au travers de la capitalisation de témoignages de professionnels du domaine et d’une exploitation de la littérature disponible
Business strategies for sustainable meat consumption: multistage investigation into profiling environmentally conscious consumer segments
No full textFNEGE 3, ABS 3International audienceThe market for sustainable meats (SMs) is now in its initial phases of developing into a commercial enterprise. To ensure the success of sustainable alternative meats, it is necessary to have a comprehensive understanding of the consumer motivations, barriers and perceived values associated with these meats. Both the Theory of Consumption Values (TCV) and the Behavioural Reasoning Theory (BRT) were used in this research project to identify and characterise various potential consumer segments that exist within the SMs market. Using the mall intercept technique, data were collected from four different cities. Study 1 comprised 458 participants, whereas Study 2 included 463. Consumers were divided into 6 cluster groups using sequential cluster analysis, considering a variety of consumption values as well as reasons for and against adopting SMs. The findings indicate that the consumers in the different clusters are distinguished by distinct combinations of perceived consumption values and reasons for or against the consumption of SMs. The findings of the study have significant implications for producers, marketers and advocates for sustainable meats, as they can help them build product, marketing and positioning strategies that are specifically targeted to each consumer segment
Multiple, diverse endogenous giant virus elements within the genome of a brown alga
No full textInternational audienceEndogenous viral elements (EVEs) have been found in diverse eukaryotic genomes. These elements are particularly frequent in the genomes of brown algae (Phaeophyceae) because these seaweeds are infected by viruses (Phaeovirus) of the phylum Nucleocytoviricota (NCV) that are capable of inserting into their host’s genome as part of their infective cycle. A search for inserted viral sequences in the genome of the freshwater brown alga Porterinema fluviatile identified seven large EVEs, including four complete or near-complete proviruses. The EVEs, which all appear to have been derived from independent insertion events, correspond to phylogenetically diverse members of the Phaeovirus genus and include members of both the A and B subgroups of this genus. This latter observation is surprising because the two subgroups were thought to have different evolutionary strategies and were therefore not expected to be found in the same host. The EVEs contain a number of novel genes including a H4 histone-like sequence but only one of the EVEs possesses a full set of NCV core genes, indicating that the other six probably correspond to nonfunctional, degenerated viral genomes. The majority of the genes within the EVEs were transcriptionally silent and most of the small number of genes that showed some transcriptional activity were of unknown function. However, the existence of some transcriptionally active genes and several genes containing introns in some EVEs suggests that these elements may be undergoing some degree of endogenization within the host genome over time
Méthodes computationnelles basées sur l'apprentissage profond pour la prédiction des structures 3D d'ARN en
No full textRNAs are, like proteins, biological molecules that play essential roles at various stages in the life of an organism and are involved in various diseases. Determining their structure, especially 3D, is essential to understand their function better. However, this is a complex problem to solve either by experimental methods (crystallography, NMR), which are very costly in time and money or by computational methods. Recently, Google DeepMind proposed a method called AlphaFold, for the prediction of the 3D structure of proteins based on deep learning, which revolutionized the field by showing a high outperformance compared to the state-of-art. However, RNA and protein molecules differ significantly in structure and dynamics, making it non-trivial to apply protein-based methods directly to RNA. AlphaFold, AlphaFold~2, as well as AlphaFold 3, its new version that also predicts RNA 3D structure, rely heavily on multiple sequence alignments (MSAs) as input, which are expensive to compute and not always available, especially for RNAs. Indeed, a lot of RNAs have unknown families, which prevents the generalisation of methods based on MSA.In this thesis, we aim to get ride of the MSA information for the prediction of RNA 3D structures. We seek to develop methods to predict RNA 3D structures from sequence information only. For this, we leverage deep learning methods and particularly language-based models to map sequences to structure features. By using language-based models pretrained on a large set of RNA sequences, we can learn RNA structural features and then predict the 3D structure.The work in this thesis is separated into three main contributions. The first, called RNAdvisor, is a tool that wraps the state-of-the-art RNA 3D structure assessment tools to comprehensively evaluate RNA 3D structures, both with and without experimental references. The second contribution, State-of-the-RNArt, is a benchmark of the state-of-the-art RNA 3D structure prediction methods, highlighting current methods' limitations and challenges. It is followed by a more detailed analysis of the limitations of AlphaFold 3.The third contribution, RNA-TorsionBERT, is a deep learning method that predicts the torsion angles of RNA 3D structures from the sequence, which are an important feature of RNA 3D structures. It leverages a language-based model to map sequences to structure features. It is extended to a new scoring function, TorsionBERT-MCQ, that can assess the quality of RNA 3D structures in torsional space. This work is a step towards the development of deep learning methods for RNA 3D structure prediction, using only sequence information and not relying on costly multiple-sequence alignments.Les ARN sont, comme les protéines, des molécules biologiques jouant des rôles essentiels à divers stades de la vie d'un organisme et impliqués dans diverses maladies. Déterminer leur structure, notamment 3D, est un enjeu essentiel pour mieux comprendre leur fonction. Or il s'agit d'un problème difficile à résoudre, aussi bien par des méthodes expérimentales (cristallographie, RMN) qui sont très coûteuses en temps et en argent, que computationnelles. Récemment, Google DeepMind a proposé une méthode appelée AlphaFold, pour la prédiction de la structure 3D des protéines basée sur l'apprentissage profond, qui a révolutionné le domaine en montrant une efficacité des prédictions très largement au-dessus de l'état de l'art. Cependant, les molécules d'ARN et de protéines diffèrent fortement en termes de structure et de dynamique, ce qui rend non trivial l'adaptation directe des méthodes développées pour les protéines aux ARN.AlphaFold, AlphaFold 2, ainsi que sa nouvelle version AlphaFold 3, qui prédit également la structure 3D des ARN, s'appuient fortement sur les alignements de séquences multiples (MSA), qui sont coûteux à calculer et ne sont pas toujours disponibles, en particulier pour les ARN. En effet, de nombreux ARN n'appartiennent à aucune famille connue, ce qui empêche la généralisation des méthodes reposant sur les MSA. Dans cette thèse, nous visons à explorer la prédiction de la structure 3D de l'ARN sans utiliser l'information issue des alignements multiples. Nous cherchons à développer des méthodes pour prédire les structures 3D des ARN à partir uniquement de la séquence. Pour cela, nous utilisons des méthodes d'apprentissage profond, et en particulier des modèles de langage, afin de faire le lien entre les séquences et les caractéristiques structurales. En exploitant des modèles de langage préentraînés sur un grand nombre de séquences d'ARN, nous pouvons apprendre des représentations riches des caractéristiques structurales de l'ARN, et ainsi prédire leur structure 3D.Le travail de cette thèse est divisé en trois contributions principales. La première, appelée RNAdvisor, est un outil qui intègre les outils d'évaluation de la structure 3D des ARN les plus récents pour évaluer de manière exhaustive les structures 3D des ARN, avec et sans références expérimentales. La deuxième contribution, State-of-the-RNArt, est un benchmark des méthodes de prédiction de la structure 3D de l'ARN les plus récentes, mettant en évidence les limites et les défis des méthodes actuelles. Elle est suivie d'une analyse plus détaillée des limites d'AlphaFold 3, la dernière version d'AlphaFold adaptée à la prédiction de la structure 3D de l'ARN. La troisième contribution, RNA-TorsionBERT, est une méthode d'apprentissage profond qui prédit les angles de torsion des structures 3D de l'ARN à partir de la séquence. Elle s'appuie sur un modèle de langage pour mettre en correspondance les séquences avec les caractéristiques de la structure. Cette méthode est étendue à une nouvelle fonction de scoring, TorsionBERT-MCQ, qui permet d'évaluer la qualité des structures 3D de l'ARN dans l'espace des torsions.Ce travail constitue une étape vers le développement de méthodes d'apprentissage profond pour la prédiction de la structure 3D des ARN, en utilisant uniquement des informations sur la séquence et sans s'appuyer sur des alignements de séquences multiples coûteux
Revisiting the functional features of maize dispensable genes
No full textInternational audienceDispensable genes are present in a subset of individuals of a species, in contrast to core genes that are shared among individuals. This fraction of the genome is significant, and could represent up to 40% of the total gene number in maize1. Dispensable genes were found to exhibit contrasted features as compared to core genes: on average, they are shorter, expressed at lower levels, more often expressed in specific tissues or conditions and they appear to be enriched in functions associated with environmental and defense responses2,3. While dispensable genes can be absent from many individuals, they have been proposed to play a role in adaptation to local environment. To better characterize the role of dispensable genes, we have generated a pan-gene set from de novo assembled genomes of 7 maize genotypes. Using an extensive transcriptomic dataset that we generated for these 7 genotypes and B73 on 15 tissues among which 7 collected from plants grown in two watering conditions, we compared the expression patterns of dispensable and core genes. Our results shed new light on the functional characteristics of maize core and dispensable genes.1.Hufford, M. B. et al. De novo assembly, annotation, and comparative analysis of 26 diverse maize genomes. Science 373, 655–662 (2021).2.Darracq, A. et al. Sequence analysis of European maize inbred line F2 provides new insights into molecular and chromosomal characteristics of presence/absence variants. BMC Genomics 19, 119 (2018).3.Tao, Y., Zhao, X., Mace, E., Henry, R. & Jordan, D. Exploring and Exploiting Pan-genomics for Crop Improvement. Mol. Plant 12, 156–169 (2019)