HAL-IRD
Not a member yet
83947 research outputs found
Sort by
Multi-step vesicle and shear-parting evolution as a record of syn-and post-emplacement deformation and cooling dynamics of 'a'a and pahoehoe lavas
No full textWhile bubbles play an intimate role in determining lava flow dynamics, the resultant vesicle shapes and distributions provide insights into the temporal and spatial evolution of lava flow emplacement. We here examine the size and shape of voids (due to vesiculation and shearing), and relate circularity to degree of deformation, for a well-exposed lava flow at Lemptégy volcano (Chaîne des Puys, France). Using high-spatial resolution imaging and textural analysis, we establish a quantitative framework to distinguish between undeformed and deformed vesicles, and fractured vesicles and their development into shear partings. Rheological modeling, using a model that evolves as the lava cools, plus calculated strain rates and shear stresses, reveal how void space deformation evolves under changing flow regimes. Our model traces a progression from initial bubble deformation under flow-induced stress, to the formation of fractures within brittle layers as cooling advances. Highly deformed bubbles occur in still fluid lava at strain rates of >7 s -1 and viscosities of 10 4 -10 7 Pa s. However, as the lava solidifies, down-slope creep takes over with well-formed shear partings developing at high mechanical stresses. In our case, this transition began after a few hours, with flow lateral creep resulting from surface up-lift due to the emplacement of a cryptodome. This study provides a method of automatically classifying void space types that make up total rock porosity and uses the time-dependent evolution of void space types to recreate and quantify syn-and post-emplacement dynamics.</div
Uncovering diversity and abundance patterns of CO2-fixing microorganisms in peatlands
No full textInternational audienceMicroorganisms play a crucial role in the carbon (C) dynamics of peatlands — a major terrestrial C reservoir. Because of their role in C emissions, heterotrophic microorganisms have attracted much attention over the past decades. CO2-fixing microorganisms (CFMs) remained largely overlooked, while they could attenuate C emissions. Here, we use metabarcoding and digital droplet PCR to survey microorganisms that potentially fix CO2 in different peatlands. We demonstrate that CFMs are abundant and diverse in peatlands, with on average 1021 CFMs contributing up to 40% of the total bacterial abundance. Using a joint-species distribution model, we identified a core and a specific CFM microbiome, the latter being influenced by temperature and nutrients. Our findings highlight that ASV richness and community structure were direct drivers of CFM abundance, while environmental parameters were indirect drivers. These results provide the basis for a better understanding of the role of CFMs in peatland C cycle inputs
Reinventar a avaliação da pesquisa à luz da ciência aberta
No full textSince the last decade, there has been a multiplication of recommendations, institutional initiatives, and government policies aimed at promoting open science. Paradoxically, research evaluation systems — increasingly anchored in quantitative indicators of productivity and impact — constitute one of the main obstacles to the dissemination of open and collaborative research practices. By prioritizing publication metrics and citation counts, these evaluative models tend to overlook — or even discourage — practices oriented toward knowledge sharing. Conversely, there has also been a growing global movement to reform the current parameters, methods, and criteria of research assessment, which are widely seen as subsumed under a narrow set of indicators that impoverish the understanding of research activity. In light of this context, this study investigates the emerging trends and challenges involved in the reform of dominant research evaluation models, taking into account the agendas propelled by the open science movement. It examines both the critiques directed at the centrality of metrics in evaluative processes and the alternative proposals aimed at reconfiguring research evaluation in alignment with the values, principles, and practices of open science. The analysis is grounded in the understanding of assessment systems as instruments of science governance, capable of shaping the ways in which knowledge is produced, circulated, and legitimized. A relational approach is adopted — focusing on the alliances and interactions among various actors — and draws from theoretical and analytical frameworks in Information Science, in dialogue with the Philosophy of Information and Science and Technology Studies (STS). The methodology includes a scoping review inspired by the PRISMA-ScR guidelines, aimed at mapping the state of the art on the reconfiguration of research assessment in its interface with open science agendas. The review encompassed academic literature published in Portuguese, French, Spanish, and English, retrieved from six information sources: Redalyc, LA Referencia, HAL, SciELO, Web of Science, and Scopus. The findings show that practices of scientific knowledge arbitration have never been static; since the founding of the first scientific societies and academies, they have been continuously reshaped through processes of contestation and (re)negotiation. This historical perspective makes it possible to denaturalize the widespread assumption that peer review has always been the primary and self-evident means of ensuring the rigor and quality of research. The study further traces connections between the scientific communication system, science and information policies, and accountability regimes, highlighting how these elements, in articulation, laid the groundwork for the consolidation of the contemporary research assessment model. The analysis then turns to the diverse conceptions of open science, revealing important convergences between reform-oriented assessment proposals and agendas promoting scientific openness. A total of 52 mobilizing initiatives aimed at transforming the hegemonic model of research assessment were identified. Subsequently, the systematic analysis of 68 studies enabled the identification of key challenges, critiques, and alternative proposals for reforming research assessment in accordance with the principles of open science. Broadly speaking, the literature underscores the urgency of transforming assessment systems in order to recognize and reward a broader diversity of research contributions. The conceptual anchor for this transformation has increasingly coalesced around the notion of Responsible Research Assessment (RRA). Finally, the study discusses how the ongoing platformization of academic infrastructures introduces new tensions and challenges for the realization of RRA and for the framing of science as a global public good. The study concludes that the platformization of digital scientific information infrastructures — under the hegemony of major publishing and data analytics corporations — has intensified the dependence of research institutions, funding agencies, and researchers on commercial tools. While such platforms offer operational conveniences, their proprietary and opaque architectures conflict with the principles of open science and Responsible Research Assessment. In this context, the reinvention of evaluation systems requires the strengthening of public infrastructures for scientific information and communication, guided by ethical and political commitments to building a more just, inclusive, and epistemically diverse scientific system.On observe, notamment au cours de la dernière décennie, une multiplication des recommandations, des initiatives institutionnelles et des politiques publiques visant à promouvoir la science ouverte. Paradoxalement, les systèmes d’évaluation de la recherche, de plus en plus appuyés sur des indicateurs quantitatifs de productivité et d’impact, constituent l’un des principaux obstacles à la diffusion de pratiques de recherche ouvertes et collaboratives. En accordant une priorité aux métriques de publication et au comptage des citations, les modèles évaluatifs actuellement en place tendent à négliger — voire à décourager — les démarches orientées vers l’ouverture et le partage des connaissances. Dans le même temps, des initiatives appelant à une refonte des paramètres, méthodes et critères d’évaluation de la recherche gagnent en visibilité à l’échelle internationale. Ces initiatives dénoncent une dépendance excessive à un ensemble limité d’indicateurs, qui réduit la compréhension des activités de recherche et de leurs apports. Face à ce constat, ce travail propose d’analyser les tendances et les défis liés à la reconfiguration du modèle dominant d’évaluation de la recherche, à partir des agendas portés par le mouvement de la science ouverte. Il s’agit d’explorer les critiques adressées à la centralité des métriques dans les processus évaluatifs ainsi que les propositions émergentes visant à transformer l’évaluation en cohérence avec les valeurs, principes et pratiques de la science ouverte. Ce travail repose sur l’idée que le système d’évaluation constitue un instrument de gouvernement de la science, en interaction avec les dynamiques de production, de circulation et de légitimation des savoirs. Une approche relationnelle est adoptée, mettant l’accent sur les alliances et interactions entre une pluralité d’acteurs, à partir d’un cadre théorique issu des sciences de l’information, en dialogue avec la philosophie de l’information et les études en science, technologie et société (STS). Une Scoping Review, inspirée des lignes directrices PRISMA-ScR, a été menée afin de cartographier l’état des connaissances sur la reconfiguration de l’évaluation de la recherche à l’intersection des agendas de la science ouverte. Cette revue s’est appuyée sur un repérage systématique de la littérature académique et scientifique en portugais, français, espagnol et anglais, à partir de six sources d’information : Redalyc, LA Referencia, HAL, SciELO, Web of Science et Scopus. Les résultats montrent que les pratiques d’arbitrage des savoirs scientifiques, loin d’être figées, ont été historiquement reconfigurées par des processus continus de tensions et de (re)négociations, depuis la création des premières académies et sociétés savantes. Cette perspective permet de déconstruire l’idée selon laquelle l’évaluation scientifique aurait toujours été conduite par les pairs dans le seul but d’assurer la rigueur et la qualité des recherches. Dans un deuxième temps, l’analyse établit des liens entre le système de communication scientifique, les politiques de science et d’information, et les régimes d’imputabilité (accountability regimes), en mettant en évidence la manière dont ces éléments, articulés entre eux, ont contribué à stabiliser le modèle contemporain d’évaluation scientifique. L’étude identifie ensuite différentes conceptions de la science ouverte, en soulignant les convergences entre les propositions de réforme des systèmes évaluatifs existants et les stratégies visant à favoriser l’ouverture de la science. Cinquante-deux actions structurantes orientées vers la transformation du modèle hégémonique d’évaluation ont été recensées. Dans une quatrième étape, à partir de l’analyse de 68 études, les principaux défis, critiques et propositions alternatives visant à réformer l’évaluation ont été extraits, en vue de l’aligner avec les principes de la science ouverte. De manière générale, les études soulignent l’urgence de réviser les systèmes d’évaluation pour qu’ils reconnaissent et valorisent une plus grande diversité de contributions issues du travail de recherche. La notion d’évaluation responsable de la recherche (Responsible Research Assessment, ou RRA) s’affirme progressivement comme une boussole pour orienter cette transformation. Enfin, ce travail discute des tensions introduites par le processus de plateformisation académique, qui complexifie la mise en œuvre de l’évaluation responsable et remet en question l’idée de la science comme bien public mondial. Il en ressort que l’expansion des plateformes numériques d’information scientifique — dominées par de grandes entreprises éditoriales et courtiers de données — accentue la dépendance des institutions de recherche, bibliothèques, agences de financement et chercheurs vis-à-vis d’outils numériques commerciaux. Si ces plateformes apportent des solutions techniques efficaces, leur caractère propriétaire et opaque entre en contradiction avec les principes de la science ouverte et de l’évaluation responsable. Dans ce contexte, nous soutenons que la transformation des systèmes d’évaluation suppose le renforcement d’infrastructures publiques de l’information et de la communication scientifique, guidées par des engagements éthico-politiques en faveur d’un système scientifique plus équitable, inclusif et pluriel.Multiplicaram-se, especialmente desde a última década, as recomendações, iniciativas institucionais e políticas governamentais que buscam promover a ciência aberta. Paradoxalmente, os sistemas de avaliação da pesquisa, cada vez mais fundamentados em indicadores quantitativos de produtividade e impacto, têm se constituído como um dos principais entraves à disseminação de práticas de pesquisa aberta e colaborativa. Ao priorizarem métricas de publicação e contagem de citações, os modelos avaliativos vigentes negligenciam — ou desestimulam — ações orientadas à abertura e ao compartilhamento do conhecimento. Por outro lado, também vêm ganhando força, em escala global, iniciativas que propõem reformular os atuais parâmetros, métodos e critérios de avaliação da pesquisa, vistos como subsumidos a um conjunto limitado de indicadores, que reduz e empobrece a compreensão sobre as atividades de pesquisa e suas contribuições. Frente a esse cenário, busca-se investigar as tendências e desafios em torno da reformulação do modelo dominante de avaliação da pesquisa, considerando as agendas impulsionadas pelo movimento de ciência aberta, explorando tanto as críticas dirigidas à centralidade conferida às métricas nos processos avaliativos quanto as propostas emergentes que visam transformar a avaliação em consonância com valores, princípios e práticas da ciência aberta. Para tanto, parte-se da compreensão do sistema de avaliação como um instrumento de governança da ciência, capaz de moldar os modos de produção, circulação e legitimação do conhecimento. Adota-se uma abordagem relacional — que privilegia, na análise, as alianças e interações entre diversos atores —, alicerçada em referenciais teóricos e analíticos da Ciência da Informação, em sua articulação com a Filosofia da Informação e com os Estudos de Ciência, Tecnologia e Sociedade. Conduz-se, também, uma Revisão de Escopo (Scoping Review), inspirada nas diretrizes PRISMA-ScR, a fim de mapear o estado da arte sobre a reconfiguração da avaliação da pesquisa em sua interface com as agendas da ciência aberta, a partir do levantamento sistemático da produção acadêmico-científica em português, francês, espanhol e inglês — proveniente de seis fontes de informação: Redalyc, LA Referencia, HAL, SciELO, Web of Science e Scopus. Como resultados, observou-se que as práticas de arbitragem do conhecimento científico, longe de serem imutáveis, foram, desde a fundação das primeiras sociedades e academias científicas, reconfiguradas por meio de constantes processos de embates e (re)negociações. Isso permitiu desnaturalizar a noção recorrente de que a avaliação científica sempre foi conduzida pelos pares com o objetivo precípuo de assegurar o rigor e a qualidade das pesquisas. Em um segundo momento, foram traçadas conexões entre o sistema de comunicação científica, as políticas científicas e de informação e os regimes de responsabilização (accountability regimes), destacando-se como esses elementos, em associação, configuraram condições constitutivas para a estabilização do modelo contemporâneo de avaliação da ciência. Em seguida, a análise contemplou diferentes concepções de ciência aberta, evidenciando convergências entre as propostas de reforma dos sistemas vigentes de avaliação e o conjunto de diretrizes e estratégias que buscam promover a abertura da ciência. Foram elencadas 52 ações mobilizadoras voltadas a transformar o modelo hegemônico de avaliação da pesquisa. Num quarto momento, a partir da análise de 68 estudos, foram extraídos os principais desafios, críticas e propostas alternativas para reformar o atual modelo de avaliação, de modo a alinhá-lo à ciência aberta. De forma geral, os estudos ressaltam a urgência de uma mudança nos sistemas de avaliação da pesquisa, a fim de que se passe a reconhecer e a recompensar uma diversidade maior de contribuições do trabalho de pesquisa. A bússola para essa reforma vem se estabelecendo em torno da noção de Avaliação Responsável da Pesquisa (Responsible Research Assessment — RRA, no inglês). Posteriormente, discutiu-se como o processo de plataformização acadêmica introduz novas tensões e desafios para a concretização da abordagem da Avaliação Responsável da Pesquisa, e para o construto da ciência como um bem público global. Concluiu-se, por fim, que o avanço da plataformização das infraestruturas digitais de informação científica — sob a hegemonia de grandes corporações editoriais e corretora de dados (data brokers) — tem intensificado a dependência de instituições de pesquisa, bibliotecas, agências de fomento e pesquisadores em relação a ferramentas digitais comerciais. Embora ofereçam facilidades operacionais, o uso de ferramentas proprietárias e opacas colide com os princípios da ciência aberta e da Avaliação Responsável da Pesquisa. Nesse contexto, argumenta-se que a reinvenção dos sistemas de avaliação requer o fortalecimento de infraestruturas públicas de informação e comunicação científica, orientadas por compromissos ético-políticos em direção à construção de um sistema científico mais justo, inclusivo e diverso
Global phylogenetic and functional structure of rodent assemblages
No full textInternational audienceExploring the global patterns of phylogenetic and functional structure of assemblages is key to describe the distribution of biodiversity on Earth and to predict how communities and ecosystem functioning may be affected by anthropogenic pressures. Rodent communities have been studied in this regard in the past, but previous work largely focused on desert ecosystems. Here, we leveraged a large database of rodent range maps, functional traits, and phylogenetic trees to compute several metrics of functional and phylogenetic structure across > 10 000 rodent assemblages spanning all terrestrial biomes. We found that the vast majority of assemblages did not significantly differ from random association among species. Importantly, we show that the current patterns we observed can locally differ from past community structure, revealing the role humans have played in altering large‐scale biodiversity patterns. We also showed a strong scale‐dependence of our metrics and revealed a weak correlation between phylogenetic and functional structure, providing an additional line of evidence that they do not reflect the same processes of community assembly
Protein Function Prediction: Graph Neural Networks as Multimodal Aggregators of Sequences, Networks, and Ontologies
No full textNational audienceBackgroundProtein function prediction has evolved from sequence-based methods to deep learning-based multimodal approaches incorporating diverse data streams to better reflect biological complexity. Graph Neural Networks (GNNs) are increasingly used to integrate heterogeneous data due to their ability to handle different dimensionalities and scales. However, the contributions of various data modalities to prediction accuracy remain unclear.ResultsWe use a simplified deep learning framework with a single-layer attention-based GNN on the BeProf dataset to investigate the impact of different data modalities on protein function prediction. The results confirm GNNs as effective integrators of diverse data, including protein interaction networks, textual, sequence and structural information. Our findings highlight the central role of language modelembeddings and the benefits of integrating InterPro and Gene Ontology annotations. We show that structural information, while useful at the protein scale[1], is redundant with other data types in large-scale networks. Despite using a simple architecture, we reach first place on the BeProf benchmark’s Cellular Component subset, third and fifth respectively on the Biological Processes and Molecular Function subsets, outperforming popular methods such as DeepGoZero, DeepGraphGO or TALE.ConclusionThis study provides insights into the relative contributions of different data modalities, showing that sequence information remains foundational while interactome, structural, and ontology-based features offer avenues for improvement. We highlight the trade-offs and benefits of multimodal frameworks, providing guidelines for future innovations in computational biology and machinelearning for enhanced protein functional annotation
Metamodeling of a physically based pesticide runoff model with a long short-term memory approach
No full textInternational audiencePredicting the temporal variability in pesticide runoff from agricultural fields is essential for assessing the impacts of agricultural practices on surface water quality and the associated environmental risks. Richards-based models are valuable tools for this purpose, particularly in contexts characterized by high temporal variability in rainfall intensity, such as in the Mediterranean region. These methods can be applied to a wide range of operational applications at both the field and catchment scales. However, their operational capability is limited, primarily due to their high computation times. In this study, we assessed a metamodeling approach for a Richards-based model designed to simulate the hourly variability in water and pesticide runoff over a year while reducing computation times. The proposed approach is based on long short-term memory (LSTM) models and is applied to a field-scale pesticide runoff model coupling Richards and convection-dispersion equations with the uniform mixing cell concept for pesticide remobilization and an overland flow routine. The resulting metamodel accounts for meteorological conditions, compound properties, and pesticide application dates and rates. Although the LSTM architecture is not optimized, the metamodel reduces the computation time by 90% compared with that of the initial model. The metamodel shows high overall performance in simulating hourly discharges and pesticide concentrations, with Nash–Sutcliffe efficiency values of 0.99, confirming the ability of the LSTM models to handle hydrochemical time series. The results nonetheless indicate room for improvement in specific areas, such as simulating runoff occurrence or concentrations for compounds with low half-life values. Furthermore, the results encourage further consideration of adding static parameters as inputs to LSTM models to increase their generalizability. Overall, this study provides key insights for deploying methods to assess the impacts of agricultural practices on a large scale
Role of individual and population heterogeneity in shaping dynamics of multi-pathogen shedding in an island endemic bat
No full textInternational audienceUnderstanding processes driving pathogen transmission in bats is critical to prevent spillovers and emergence events. Although substantial research has addressed this topic, few studies have directly examined shedding dynamics (as opposed to serological studies) and co-infection patterns using fine-scale spatio-temporal datasets. Here, based on the monitoring of 5,714 Reunion free-tailed bats (Mormopterus francoismoutoui) in 17 roosts over 24 months, we studied the co-shedding dynamics of paramyxoviruses (PMV) and Leptospira bacteria (LEPTO) in urine, and herpesviruses (HSV) in saliva. We evidenced all year long shedding with high prevalence of all three infectious agents (37% -87%), as well as an exceptionally high level of co-shedding (59%), with both positive and negative interactions between infectious agents. Shedding patterns displayed temporal synchrony among roosts, with a peak during summer months, but were not influenced by roost size. Repeated shedding in recaptured bats supports within-host persistence, though underlying mechanisms remain to be identified. Our results also showed rapid HSV infection of juveniles (< 6 months), and suggest longer protection of juveniles by maternal antibodies for PMV and LEPTO. Reproductively-active individuals (both during the pregnancy and mating) were associated with increased PMV and LEPTO shedding, which could result from tradeoffs between reproduction and infection in both sexes, and/or an age-related bias with the progressive infection of older juveniles during reproductive periods. This study highlights the significance of persistent shedding of multiple pathogens, including bacteria, and their intricate interactions within bat populations.Understanding how human-driven ecological changes may disrupt within-host processes and influence pathogen shedding in bats will help assessing the risk of pathogen spillover from bats to other species, including humans
Modélisation de l'exposition de la biodiversité aux produits Phytopharmaceutiques (PPP) a l’échelle du paysage
No full textInternational audienceUne limite majeure de l’évaluation des risques pour les produits phytosanitaires est que les pesticides sont généralement appliqués à l’échelle de la parcelle agricole, alors que les sources d’émission et les voies de contamination se situent à l’échelle du paysage et tendent à s’accumuler au fil du temps. Les méthodes d’évaluation des risques environnementaux (ERE) utilisées pour l’autorisation de mise sur le marché reposent souvent sur une approche substance par substance à l’échelle de la parcelle, ne permettant pas de prendre en compte l’ensemble des risques liés aux mélanges de pesticides. Pour faire face à cette complexité, il est nécessaire de développer de nouvelles approches de modélisation de l’exposition permettant de refléter fidèlement la dynamique de dispersion des pesticides en tenant compte des caractéristiques écologiques et des stades de vie propres à chaque espèce, tels que les modes d’exposition, les schémas de dispersion et les déplacements d’animaux entre différents habitats. Dans ce contexte, le modèle MIPP (développer sigle), dédié au devenir des pesticides dans les agroécosystèmes viticoles permet de simuler la dispersion des concentrations de pesticides dans l’air, le sol et les eaux de surface, tout en prenant en compte des éléments paysagers clés tels que les parcelles agricoles, les réseaux hydrographiques, les types de sols et les pratiques agricoles. Ces schémas de contamination seront croisés avec un modèle individu-centré (Individual-Based Model) permettant de représenter l’utilisation de l’habitat et les comportements alimentaires de différents groupes fonctionnels d’organismes. Ce travail permettra de mieux comprendre comment la contamination par les pesticides dans les différents milieux environnementaux (air, eau, sol) interagit avec les traits biologiques des espèces sauvages pour déterminer les risques d’exposition. Cet effort contribuera à des évaluations des risques environnementaux plus réalistes, en intégrant une vision globale des facteurs écologiques et paysagers influençant la dispersion des pesticides. L’approche retenue sera présentée à l’issue d’un aperçu des outils disponibles pour atteindre cet objectif
Single-cell imaging reveals efficient nutrient uptake and growth of microalgae darkening the Greenland Ice Sheet
No full textInternational audienceBlooms of dark pigmented microalgae accelerate glacier and ice sheet melting by reducing the surface albedo. However, the role of nutrient availability in regulating algal growth on the ice remains poorly understood. Here, we investigate glacier ice algae on the Greenland Ice Sheet, providing single-cell measurements of carbon:nitrogen:phosphorus (C:N:P) ratios and assimilation rates of dissolved inorganic carbon (DIC), ammonium and nitrate following nutrient amendments. The single-cell analyses reveal high C:N and C:P atomic ratios in algal biomass as well as intracellular P storage. DIC assimilation rates are not enhanced by ammonium, nitrate, or phosphate addition. Our combined results demonstrate that glacier ice algae can optimise nutrient uptake, facilitating the potential colonization of newly exposed bare ice surfaces without the need for additional nutrient inputs. This adaptive strategy is particularly important given accelerated climate warming and the expansion of melt areas on the Greenland Ice Sheet
A non-innocent, π-extended N-heterobicyclic carbene–thiolate ligand
No full textInternational audienceA bidentate NHC-thiolate ligand based on the imidazo[1,5-a]pyridine scaffold and featuring an extended p-system is described. It is derived from a readily available zwitterionic precursor and the chemistry of the corresponding homoleptic Ni(II) complex is investigated