Université de Neuchâtel

Machine learning (ML) training data is often scattered across disparate collections of datasets, called data silos . This fragmentation poses a major challenge for data-intensive ML applications: integrating and transforming data residing in different sources demand a lot of manual work and computational resources. With data privacy constraints, data often cannot leave the premises of data silos; hence model training should proceed in a decentralized manner. In this work, we present a vision of bridging traditional data integration (DI) techniques with the requirements of modern machine learning systems. We explore the possibilities of utilizing metadata obtained from data integration processes for improving the effectiveness, efficiency, and privacy of ML models. Towards this direction, we analyze ML training and inference over data silos. Bringing data integration and machine learning together, we highlight new research opportunities from the aspects of systems, representations, factorized learning, and federated learning.

Siderite (FeCO 3 ) is an important reservoir of mineral-bound ferrous iron in non-sulfidic, reducing soils and sediments. It is redox sensitive, and its oxidation may facilitate the reduction of a range...

Although it is well established that parenthood affects employment decisions within households, less is known about how this effect varies across couple types in different countries. Using difference-in-differences analysis with propensity score matching and multilevel modeling with cross-level interactions, this study explores heterogeneity in the effect of parenthood on couples’ division of paid labor by prebirth relative earning power in different European contexts. The results show that the decline in the female share of couples’ paid working hours after parenthood is stronger in male main-earner couples than in equal-earner or female main-earner couples. Our cross-national findings demonstrate that institutional and cultural factors influence couples’ postparenthood employment arrangements, with the magnitude of these moderating effects dependent on the couple's prebirth relative earning pattern. Longer paid maternity and family leave for mothers exacerbates the parenthood effect on couples’ division of paid labor, whereas greater childcare provision weakens this effect. However, these policy impacts are observed only among male main-earner couples, which can be attributed to differences in opportunity costs and bargaining power across couple types. Egalitarian gender norms weaken the effect of parenthood on the paid labor division between partners regardless of the breadwinner couple type before childbirth.

Human language relies on a rich cognitive machinery, partially shared with other animals. One key mechanism, however, decomposing events into causally linked agent–patient roles, has remained elusive with no known animal equivalent. In humans, agent–patient relations in event cognition drive how languages are processed neurally and expressions structured syntactically. We compared visual event tracking between humans and great apes, using stimuli that would elicit causal processing in humans. After accounting for attention to background information, we found similar gaze patterns to agent–patient relations in all species, mostly alternating attention to agents and patients, presumably in order to learn the nature of the event, and occasionally privileging agents under specific conditions. Six-month-old infants, in contrast, did not follow agent–patient relations and attended mostly to background information. These findings raise the possibility that event role tracking, a cognitive foundation of syntax, has evolved long before language but requires time and experience to become ontogenetically available.

The field of evolutionary and speciation genomics has been revolutionised by the ubiquity and availability of genomic data even for non-model organisms. The capability to sequence long-fragment DNA has particularly spurred trans-national initiatives to generate publicly available chromosome-resolved reference genomes across the Tree of Life. Initiatives such as the Darwin Tree of Life (The Darwin Tree of Life Project Consortium et al. 2022) or the European Genome Atlas (ERGA; Mazzoni, Ciofi, and Waterhouse 2023) enable researchers around the globe to address unresolved questions and pursue novel lines of research. The role of chromosomal rearrangements (CRs) in driving evolution has been a long-standing question in evolutionary biology (Berdan et al. 2023; Dobzhansky and Sturtevant 1938; King 1995; Robertson 1916; Wellenreuther and Bernatchez 2018; White 1978). CRs comprise an array of rearrangements that reorganise the linear sequence of the genome, ranging from local structural variants (SVs) such as inversions or duplications, to large-scale karyological changes, including chromosomal fusions and fissions (Berdan et al. 2023; Lucek et al. 2023). Current genomic data have already highlighted that CRs are much more common and diverse across taxa than previously thought (e.g., Damas et al. 2022; Weissensteiner et al. 2020). While CRs can now more easily be mapped within and across taxa, their potential role for evolution and species diversification has often remained enigmatic, and theoretical explorations exist for only a few types of CR (Berdan et al. 2023). At the dawn of broad genome availability for model and non-model organisms alike, our special issue aims to provide a genomic update on the evolutionary impact of various types of CRs. Specifically, our special issue asks what causes the evolution and establishment of CRs and whether these causes differ among taxa? Are CRs randomly distributed across the genome, and do they cause other chromosomal rearrangements? How can CRs promote diversification and how do they potentially lead to speciation? Is the evolutionary impact of CRs the same among different types of CRs? Finally, do CRs have a different evolutionary impact depending on whether autosomes or sex chromosomes are involved?

Four Gram-stain-negative bacterial strains, CS20T, AUT15.5T, XENO-11T, and CCN3.3T, isolated from Steinernema entomopathogenic nematodes, were found to represent novel species within the genus Xenorhabdus (Gammaproteobacteria, Morganellaceae). In this study, we described these new species using whole-genome phylogenomic reconstructions, sequence identity values from core genome sequences, and phenotypic characterization. Phylogenetic reconstructions based on 16S rRNA gene sequences showed that: (i) strain CS20T is closely related to X. stockiae DSM 17904T, (ii) strain AUT15.5T is closely related to X. budapestensis DSM 16342T, (iii) strain XENO-11T is closely related to X. khoisanae DSM 25463T, and (iv) strain CCN3.3T is closely related to X. griffiniae DSM 17911T. The 16S rRNA gene sequence similarity value between strain CS20T and X. stockiae DSM 17904T is 97.8%, between strain AUT15.5T and X. budapestensis DSM 16342T is 98.1%, between strain XENO-11T and X. khoisanae DSM 25463T is 97.8%, and between strain CCN3.3T and X. griffiniae DSM 17911T is 98.6%. Phylogenomic reconstructions using whole-genome sequences showed that: (i) strain CS20T is closely related to X. stockiae DSM 17904T and X. innexi DSM 16336T, (ii) strain AUT15.5T is closely related to X. indica DSM 17382T, (iii) strain XENO-11T is closely related to X. khoisanae DSM 25463T, and (iv) strain CCN3.3T is closely related to X. griffiniae DSM 17911T. The digital DNA-DNA hybridization (dDDH) and average nucleotide identity (ANI) values between strains CS20T, AUT15.5T, XENO-11T, and CCN3.3T and the type strains of their more closely related species are below the 70% and the 95–96% divergence thresholds, respectively, used for prokaryotic species delineation. Hence, we propose the following four new species: Xenorhabdus bharatensis sp. nov. (the type strain is CS20T=CCM 9320T=CCOS 2070T), X. entomophaga sp. nov. (the type strain is XENO-11T=CCM 9389T=CCOS 2111T), X. siamensis sp. nov. (the type strain is AUT15.5T=CCM 9405T=CCOS 2116T), and X. thailandensis sp. nov. (the type strain is CCN3.3T=CCM 9406T=CCOS 2115T). The following biochemical tests may be useful for differentiating the novel species from their more closely related taxa: acetoin production, arginine dihydrolase, citrate utilization, gelatinase, glucose oxidation, indole production, and tryptophan deaminase. Our study contributes to a better understanding of the biodiversity and phylogenetic relationships of entomopathogenic bacteria associated with insect parasitic nematodes.

Global warming is changing plant communities due to the arrival of new species from warmer regions and declining abundance of cold‐adapted species. However, experimentally testing predictions about trajectories and rates of community change is challenging because we normally lack an expectation for future community composition, and most warming experiments fail to incorporate colonization by novel species. To address these issues, we analyzed data from 44 whole‐community transplant experiments along 22 elevational gradients across the Northern Hemisphere. In these experiments, high‐elevation communities were transplanted to lower elevations to simulate warming, while also removing dispersal barriers for lower‐elevation species to establish. We quantified the extent and pace at which warmed high‐elevation communities shifted towards the taxonomic composition of lower elevation communities. High‐elevation plant communities converged towards the composition of low‐elevation communities, with higher rates under stronger experimental warming. Strong community shifts occurred in the first year after transplantation then slowed over time, such that communities remained distinct from both origin and destination control by the end of the experimental periods (3‐9 years). Changes were driven to a similar extent by both new species colonization and abundance shifts of high‐elevation species, but with substantial variation across experiments that could be partly explained by the magnitude and duration of experimental warming, plot size and functional traits. Our macroecological approach reveals that while warmed high‐elevation communities increasingly resemble communities at lower elevations today, the slow pace of taxonomic shifts implies considerable colonization and extinction lags, where a novel taxonomic composition of both low‐ and high‐elevation species could coexist for long periods of time. The important contribution of the colonizing species to community change also indicates that once dispersal barriers are overcome, warmed high‐elevation communities are vulnerable to encroachment from lower elevation species.

Soils play a central role in global biogeochemical cycles and host plants and invertebrates whose engineering activities mainly occur in the upper soil layers, the so-called humipedon. This latter is the critical zone where most chemical, physical, and biological ecosystem processes arise. This chapter overviews the main definitions and concepts (diagnostic horizon, humus system, humus form) and functional approaches to humipedon intimately linked to ecosystem engineers, mainly plant roots and earthworms. Biological activities are decisive in integrating organic matter and forming soil aggregates and galleries. We highlight the relevance of studying humus forms based on field observations in various environments (grasslands, forests, mountains, floodplains) and pioneer ecosystems and underline that more research is needed on the role played by earthworms in the evolution of terrestrial biomes.

La Draga (Banyoles, Spain) is one of the most relevant Early Neolithic site in southern Europe and the Mediterranean region, among other reasons, for its excellent preservation of organic archaeological materials in waterlogged conditions. The site corresponds to a lake dwelling of the early farmers in the region. The goal of this paper is to present the current results of the dendrochronological analysis of the piles and horizontal woods and the ¹⁴ C dates. In order to understand the relation of the wooden piles with the occupation phases documented, a Bayesian model has been built including the stratigraphy, dendrochronological and ¹⁴ C data; moreover, a wiggle-matching of the dendrochronology and ¹⁴ C dates of the piles has also been developed. The dendrochronological results show three phases of tree felling, a single main construction event and establish a minimum duration of the wooden constructions at la Draga of 27 years. Radiocarbon dating combined with dendrochronology confirms these results and dates the construction event around 5310 cal BC.

We performed a series of hydraulic stimulations at 1.1 km depth in the Bedretto underground laboratory, Switzerland, as part of an overall research strategy attempting to understand induced seismicity on different scales. Using an ultra‐high frequency seismic network we detect seismic events as small as Mw < −4, revealing intricate details of a complex fracture network extending over 100 m from the injection sites. Here, we outline the experimental approach and present seismic catalogs as well as a comparative analysis of event number per injection, magnitudes, b‐values, seismogenic index and reactivation pressures. In our first‐order seismicity analysis, we could make the following observations: The rock volume impacted by the stimulations in different intervals differs significantly with a lateral extent from a few meters to more than 150 m. In most intervals multiple fractures were reactivated. The seismicity typically propagates upwards toward shallower depth on parallel oriented planes that are consistent with the stress field and seem to a large extent associated with preexisting open fractures. This experiment confirms the diversity in seismic behavior independent from the injection protocol. The overall seismicity patterns demonstrate that multi‐stage stimulations using zonal isolation allow developing an extended fracture network in a 3D rock volume, which is necessary for enhanced geothermal systems. Our stimulations covering two orders of magnitude in terms of injected volume will give insights into upscaling of induced seismicity from underground laboratory scale to field scale.

In monoculture-dominated landscapes, recovering biodiversity is a priority, but effective restoration strategies have yet to be identified. In this study, we experimentally tested passive and active restoration strategies to recover taxonomic, phylogenetic, and functional diversity of woody plants within 52 tree islands established in an oil palm landscape. Large tree islands and higher initial planted diversity catalyzed diversity recovery, particularly functional diversity at the landscape level. At the local scale, results demonstrated that greater initial planting diversity begets greater diversity of native recruits, overcoming limitations of natural recruitment in highly modified landscapes. Establishing large and diverse tree islands is crucial for safeguarding rare, endemic, and forest-associated species in oil palm landscapes.

In times of global change, high temperatures can increase the negative effects of pesticides and other stressors. The goal of this study was to evaluate, under controlled laboratory conditions, the effect of a moderate increase in temperature in combination with ivermectin (an antiparasitic medication used in cattle that is excreted in dung), an herbicide, and parasitic pressure, on the reproductive success, development time and adult survival of dung beetles Euoniticellus intermedius. Whereas high temperature increased the number and proportion of emerged offspring, it had synergistic negative effects in combination with the ivermectin, herbicide and parasite treatments. Moreover, high temperature in combination with ivermectin and with parasitism caused a synergistic increase of adult offspring mortality and, in combination with the herbicide, it synergistically accelerated development. These results indicate that high temperatures can enhance the negative effects of other stressors and act synergistically with them, harming dung beetles, a group with high ecological and economic value in natural and productive ecosystems. Although adult sex ratio was not affected by experimental treatments, contrasting responses were found between males and females, supporting the idea that both sexes use different physiological mechanisms to cope with the same environmental challenges. The effects that combined stressors have on insects deepen our understanding of why we are losing beneficial species and their functions in times of drastic environmental changes.

Research has shown that soil-borne beneficial microorganisms can enhance plant growth, productivity, and resistance against pests and pathogens and could thus serve as a sustainable alternative to agrochemicals. To date, however, the effect of soil-beneficial microbes under commercial crop production has been little assessed. We here investigated the effect of root inoculation with nine well-characterized bacterial and fungal strains and two consortia on tomato performance under intensive tomato crop management practices. We measured the impact of these root inoculations on plant growth, fruit quality, yield, and pest and pathogen incidence. While most microbial strains showed weak effects, we found that the fungal strains Trichoderma afroharzianum T22 and Funneliformis mosseae significantly increased marketable tomato yield. Moreover, we found that inoculation with most of the fungal strains led to a significant reduction in the incidence of the devastating leaf-mining pest Tuta absoluta , while this effect was not observed for bacterial inoculants. In addition, we found that microbial inoculations did not impact the incidence of introduced natural pest enemies, supporting their compatibility with well-established integrated pest management strategies in horticulture. In summary, the observed general positive effects of soil microbes on tomato yield and resistance reinforce the move toward broader adoption of microbial inoculants in future crop production, ultimately improving agricultural sustainability.

Mating between closely related individuals can result in a reduction in offspring fitness, known as inbreeding depression. Here, we investigate whether breeding with close relatives affects the reproductive output of parents and the development of their offspring in Damaraland mole-rats (Fukomys damarensis), a cooperatively breeding species where females avoid mating with familiar individuals. By cross-fostering litters of pups soon after birth, we were able to form breeding pairs from full siblings that were reared apart. We compared the reproductive output of these sibling pairs and the survival and growth of their pups with that of unrelated pairs over a period of 4 years. The litter sizes and interbirth intervals of sibling pairs did not differ from those of unrelated pairs, but the growth and survival of inbred offspring were lower, showing that breeding between close relatives is associated with substantial fitness costs. This study suggests that inbreeding depression is an important driver of the extreme reproductive skew observed in social mole-rats. Studies of the costs of inbred matings are now needed in similar species, such as naked mole-rats (Heterocephalus glaber), where captive females more commonly breed with close relatives, to determine whether these costs are lower than in Damaraland mole-rats.

Social organisation of eusocial insects requires efficient communication among conspecifics, involving various signals. Among them, Cuticular Hydrocarbons Compounds are used like chemical signals for recognition processes. These semiochemical compounds, which can vary qualitatively and quantitatively, form an individual chemical signature carrying identity of each congeners which contribute to the social cohesion of the colony members. In this study, we analysed the chemical signature of workers of the eusocial and invasive Vespidae species, the Yellow-legged hornet, Vespa velutina nigrithorax. The chemical communication system between hornets’ workers is relatively unknown and their social organisation poorly documented. However, a strong chemical heterogeneity between castes and colonies have been previously identified in the Yellow-legged hornet, suggesting a possible chemical diversity between workers. Our results showed a strong chemical heterogeneity mainly explained by their colonial origin, as previously described, but also by their behaviour at a given time. In this study, four behaviours have been reported in the field and could be assigned to a workers’ sub-caste: animal foragers, builders, defenders and material foragers. A chemical separation of individuals into two groups have been observed, where animal foragers exhibit a clear separation of their chemical profiles compared to their counterparts. Also, animal foragers had more alkenes and fewer branched alkanes than the other workers. This exploratory study demonstrates that workers of this invasive hornet species present different cuticular profiles, probably used in both inter and intra-specific recognition phenomena. This is therefore a first step towards understanding the chemical communication involved in the social organisation of hornet workers.