University of Avignon

Bioinoculants are increasingly used in land restoration programs to alleviate environmental stress during tree establishment and to enhance vegetation recovery in degraded ecosystems. However, their effects on native soil microbiota remain insufficiently understood, particularly in highly vulnerable environments, limiting their large-scale application. In salt-affected lands of Senegal, restoration strategies rely on Casuarinaceae species inoculated with a multi-kingdom bioinoculant composed of salt-tolerant arbuscular mycorrhizal fungi and nitrogen-fixing bacteria. While this strategy successfully enhanced tree growth and understory vegetation, its consequences for soil microbial diversity and functions have never been assessed. The objective of this study was to assess whether bioinoculation alters soil microbiota and to determine the relative influence of bioinoculation compared with salinity and host plant identity. We hypothesized that bioinoculation reshapes both soil bacterial and fungal microbial communities, with implications notably on soil nutrient cycling, but that these effects are constrained by salinity and Casuarinaceae species. Our results show that bioinoculation reduced overall fungal diversity without affecting rare taxa, whereas bacterial diversity was primarily driven by salinity and host plant identity, with effects largely confined to dominant bacterial groups. Functional predictions revealed marked shifts, with bioinoculation associated with a decrease in potential bacterial pathogen guilds, while increasing salinity promoted potential fungal pathogen guilds. Bacterial-mediated nitrogen cycling responded jointly to salinity and host plant identity, whereas nitrogen-fixing bacterial taxa were specifically promoted by bioinoculation. Overall, our findings demonstrate that bioinoculation can significantly modify soil microbial diversity and predicted functions in salt-affected soils, but that its effects may be overridden by strong environmental constraints and host plant effects. These results highlight ecological trade-offs associated with bioinoculant use and emphasize the need to integrate soil microbial responses into sustainable salt-affected land restoration strategies.

The study presents the synthesis of terpene-based novel biosurfactants derived from sodium and choline salts of citronellic acid (SC, CC) and citronellyl sulfate (SCS, CCS). Surface tension (ST) measurements, dynamic light scattering (DLS), and small-angle neutron scattering (SANS) were used to examine their colloidal properties and thermodynamic parameters. The findings suggest that hydrophobic tail architecture, headgroup chemistry, and counterion identity significantly influence the micellization behavior of citronellol-based amphiphiles. Sulfate-containing surfactants (SCS and CCS) had lower critical micellar concentrations (CMCs) of 134 +/- 9.4 and 88.8 +/- 4.3 mM, respectively, than carboxylate analogs (SC and CC; similar to 200 mM). SANS analysis revealed spherical micelles with radii of 0.8-0.9 and 1.1-1.3 nm for SCS and CCS, respectively. DLS measurements further validated the presence of predominant nanoscale aggregates (< 5 nm) with reduced polydispersity, especially at higher concentration. The zeta potential data indicated that the choline counterion exhibited partial charge screening compared to sodium. All surfactants showed very low cytotoxicity (IC50 = 1149-1400 mu g/mL) against RAW264.7 macrophage immune cells and biocompatibility similar to that of lauryl glycoside and Tween 20 in HEK-293 cells. Initial studies reveal their potential as safe insect repellents and as a part of crop-protection formulations.

Simultaneous speech translation requires translating source speech into a target language in real-time while handling non-monotonic word dependencies. Traditional approaches rely on supervised training with word-level aligned data, which is difficult to collect at scale and thus depends on synthetic alignments using language-specific heuristics that are suboptimal. We propose Hibiki-Zero, which eliminates the need for word-level alignments entirely. This fundamentally simplifies the training pipeline and enables seamless scaling to diverse languages with varying grammatical structures, removing the bottleneck of designing language-specific alignment heuristics. We first train on sentence-level aligned data to learn speech translation at high latency, then apply a novel reinforcement learning strategy using GRPO to optimize latency while preserving translation quality. Hibiki-Zero achieves state-of-the-art performance in translation accuracy, latency, voice transfer, and naturalness across five X-to-English tasks. Moreover, we demonstrate that our model can be adapted to support a new input language with less than 1000h of speech. We provide examples, model weights, inference code and we release a benchmark containing 45h of multilingual data for speech translation evaluation.

A lack of standardised sampling protocols prevents functional traits from expressing their full potential to advance plant ecology, biogeography, and evolutionary biology. Handbooks providing protocols for standardised measurements of plant functional traits allow researchers to tackle large-scale ecological questions but have traditionally focused on vegetative traits such as leaves, stems and roots. This handbook provides standardised protocols for 58 reproductive traits of flowers (10 traits), fruits (6 traits), seeds (36 traits) and seedlings (6 traits). It is the first effort to standardise sampling for relevant reproductive traits to better understand processes, such as pollination, frugivory, seed dispersal, seed longevity, germination, and seedling establishment. The protocols were designed to embrace the global diversity of ecological contexts experienced by flowers, fruits, seeds, and seedlings and incorporate methods for temperate to tropical, dry to moist and fire-prone to fire-sensitive ecosystems. We offer general guidelines for sampling, storing, and processing reproductive traits. Before laying out the protocol, we briefly describe each trait functionality, trade-offs, and sources of variability to give a broad context. Standardised protocols to estimate reproductive plant traits will unlock the full potential of plants to mitigate land-use and climate-change impacts, and support restoration of degraded ecosystems.

investigate whether the multiwavelength variability of quasar 3C 273 can be described in terms of nonextensive statistical mechanics. Using a dataset of 39 light curves spanning from radio to X-rays collected by the Integral Science Data Centre (ISDC) database, we apply the q-Gaussian formalism to determine the entropic index q for each spectral band. Our analysis reveals a nontrivial dependence of q on energy, with two prominent increases from radio to mm/submm and hard X-ray domains. The first linear trend is likely associated with synchrotron flares and thermal emission from hot dust, while the second may reflect inverse Compton processes in a turbulent jet environment. However, the second linear trend in the hard X-ray range lacks statistical robustness due to limited data or large measurement uncertainties. Interestingly, the q values show no clear correlation with the fractional variability amplitude Fvar, suggesting that q captures a different aspect of the source's dynamical behavior. These findings offer new insights into how different emission mechanisms and physical regions in AGNs may manifest distinct degrees of statistical complexity. Copyright c 2026 EPLA All rights, including for text and data mining, AI training, and similar technologies, are reserved.