Systematic Entomology

The Asparagaceae is an economically and ecologically important angiosperm (monocot) clade, containing iconic garden plants such as asparagus, hyacinth and agave, and arid-adapted keystone species like Joshua tree and dragon blood tree. The approximately 2,500 species in the family are spread across seven subfamilies: Agavoideae, Aphyllanthoideae, Asparagoideae, Brodiaeoideae, Lomandroideae, Nolinoideae and Scilloideae. Relationships among the seven subfamilies are generally unresolved and the sources of gene tree-species tree discordance are largely unknown. Investigating the biology underlying such discordance requires a robust sampling of both taxa and molecular markers, due to the extremely short branches that form the backbone of the Asparagaceae phylogeny. Using targeted sequencing of 1,726 low-copy nuclear genes, and a nearly complete sampling of genera, we interrogate relationships among all seven Asparagaceae subfamilies and investigate the underlying sources of discordance across the phylogeny. Analyses using different phylogenomic approaches (multispecies coalescence vs. nuclear gene supermatrix vs. chloroplast gene supermatrix) each revealed extensive discordance across the Asparagaceae backbone, which may be associated with rapid lineage diversifications that occurred in Africa shortly following the origin of the family ~49 Ma. Specifically, we found that two branches, representing ancestors of Aphyllanthoideae and Nolinoideae, may be responsible for much of the backbone discordance. The placement of Aphyllanthoideae, represented by one extant taxon (Aphyllanthes monspeliensis), is controversial with discordant topologies across many studies. However, multispecies coalescence-based analyses from this study strongly support a sister relationship between Aphyllanthoideae and a larger Agavoideae + Scilloideae + Brodiaeoideae clade. According to gene tree quartet frequencies and ABBA-BABA tests, we cannot reject incomplete lineage sorting (ILS) as the main source of discordance for the placement of Aphyllanthoideae; our results leave open the possibility of an ancient reticulation involving an ancestor of Aphyllanthoideae and an extinct basal Asparagaceae lineage. Regarding Nolinoideae, we found that conflicting support for two possible bifurcating clades (Asparagoideae + Nolinoideae vs. Asparagoideae + Lomandroideae) was associated with skewed support for alternative gene tree topologies, suggesting that phenomena other than ILS may explain the observed discordance. A quartet-based ABBA-BABA analysis, used to test for signatures of introgression, revealed a putative ancient hybridization event between two divergent, ancestral lineages which may have contributed to the origin of the Nolinoideae clade in Africa ~45 Ma. In sum, this study elucidates the long-troublesome backbone of the Asparagaceae phylogeny and advances our understanding of the evolutionary and biogeographical origins of all extant subfamilies. More broadly, this study expands our understanding of the various sources of gene tree discordance that may arise from natural processes associated with rapid lineage radiations across the tree of life.

Premise of the study: Doryopteris J.Sm. is a genus of dry-adapted ferns found widely across the southern hemisphere and the tropics. In the Hawaiian Islands, the genus is represented by four endemic species and one putative hybrid. Two of the endemic species (D. decora and D. decipiens) are widespread across all major islands, while two (D. angelica and D. takeuchii) are federally and state-listed endangered single-island endemics.We aimed to resolve the evolutionary relationships and biogeographic history of Hawaiian Doryopteris. Methods: We used a target-capture approach and the GoFlag 408 flagellate land plant probe set to sequence more than 400 nuclear loci for 40 Hawaiian and 29 non-Hawaiian individuals in Doryopteris and the related genera Lytoneuron and Ormopteris. After processing the sequences into alignments and single nucleotide polymorphism sites, we performed phylogenetic inference and population-level principal component analyses. Key results: We infer that Hawaiian Doryopteris are monophyletic and are nested within the pantropical species D. concolor. Within Hawai'i, clades and clusters of individuals correspond to geographical, geological, and climatological factors, more so than species identifications, with a few exceptions. We posit that Hawaiian Doryopteris represents an incipient radiation at the very earliest stages of divergence, and that budding speciation has led to non-monophyly of several taxa. Conclusions Our inference using genome-wide sequencing data sheds light on the complex evolutionary history of Hawaiian Doryopteris. Overall, the phylogeny of Hawaiian Doryopteris shows strong links with geographical, geological, and climatological processes and little with morphological evolution.

Kashmir voles (Hyperacrius) are poorly studied fossorial arvicolines of the southern Himalayan slopes, with two species recognized. We sequenced Cytb from voles collected in northern Pakistan and combined these data with existing GenBank sequences, yielding 194 tips spanning nearly all arvicoline genera. We analyzed five data sets with maximum likelihood and compared external measurements among species using principal component and linear discriminant analyses. We recovered five well-supported clades of Hyperacrius separated by ~12-14% p-distance, which we treat as distinct species. We recognize clade 1 as H. wynnei, and populations referred to H. w. traubi are genetically indistinguishable from H. w. wynnei. We recognize H. zygomaticus, new rank (clade 2), and tentatively allocate clade 5 to H. brachelix. Clades 3 and 4 likely represent undescribed species. Morphometric analyses of external measurements clearly distinguish H. wynnei but not clades 3 and 4 (overall classification accuracy 70.3%). Phylogenetic analyses including nuclear genes placed Hyperacrius as sister to a clade containing Microtini, Arvicola, Lagurini, Dinaromys, and Ellobiusini, whereas mitochondrial analyses weakly allied it with Ellobiusini. Its traditional placement in Clethrionomyini is thus unsupported, although its exact position remains unresolved. Alticola is nonmonophyletic, with A. macrotis sequences nested within Clethrionomys, likely reflecting mitochondrial capture.

Speciation can be a slow process for long-lived organisms with large populations due to persistent incomplete lineage sorting, while hybridisation can further obscure species boundaries through admixture. In eucalypts, there is tremendous morphological and environmental variation with almost a thousand described species, subspecies, and hybrid taxa falling into multiple subgenera and intermediate sections. Multilocus molecular studies have generated concatenated species trees supporting deeper divergence between the Corymbia + Blakella + Angophora clade with the rest of eucalypts. However, the proportion of gene trees that recover the concatenated tree topology, including the taxonomic relationships among individual subgenera and sections, remains underappreciated. In this study, we assessed the predictive power of current eucalypt taxonomy - represented as a concatenated tree - in explaining topological variation across gene trees. To do this, we first extracted 1,187 BUSCO loci from short-read data of 701 samples representing roughly 500 described species using CAPTUS. We then inferred individual gene trees and a concatenated species tree using IQ-TREE2. Finally, we calculated the concordance factors for every described subgenus and section on the concatenated tree, and developed a taxonomic consistency test that tracks the taxonomic placement of each sample across gene trees. We show that, even though most taxonomic groups are monophyletic on the concatenated tree with 100% bootstrap support, several groups exhibit very low concordance, indicating that most gene trees do not recover their relationships as represented by the concatenated tree. These results highlight important considerations for current eucalypt taxonomic classification and provide a phylogenomic inference pipeline that can be applied not only to future eucalypt studies, but also to other evolutionarily complex groups of taxa.

Motivation: Tree sets appear naturally when performing model based phylogenetic analyses. When tracking host-parasite co-evolution, testing sensitivity of different model assumptions, or even just checking convergence of individual runs of the same analysis, the question arises: are a given pair of tree sets the same? And if they are not the same, how do they differ? Results: We provide DensiTree 3, an interactive tool for inspecting pairs of sets of trees that allow visualising both tree sets using a single panel. Incorporation of CCDs provide better summary trees, taxon orderings and extra functionality like rogue detection, making it easier to interpret tree samples. Furthermore, a clade comparison graph allows inspection of differences in clade support and clade heights linked to the double-DensiTree panel. Together, these visualisations of tree sets make it easy to identify potential differences in the tree sets and interactively inspect these differences. Furthermore, recent advances in conditional clade distributions (CCDs) to represent tree distributions have been incorporated in DensiTree 3, and the most visible part of the user interface has been updated for a more modern look. Availability: DensiTree 3 is freely available from https://cs.auckland.ac.nz/~remco/DensiTree/. The program is licensed under GPL and source code is available via the link above.

Three-dimensional imaging is rapidly increasing the scale of biological morphometric datasets, but establishing automated anatomical correspondence remains a major bottleneck. Conventional point-cloud registration methods transfer landmarks efficiently, yet poor initialization or high morphological variation can produce geometrically plausible but biologically implausible deformations. Here, we developed a statistical-shape-model (SSM)-regularized coherent point drift (CPD) estimator for automated morphometric landmark transfer. It replaces the unconstrained non-rigid CPD update with optimization in SSM coefficient space and regularizes coefficients by inverse population covariance, so deformations along low-variance biological modes are penalized more strongly. We compared pose-matched conventional CPD, SSM-CPD, and SSM-CPD with refinement in leave-one-out atlas experiments on laboratory mouse skulls (N=249), primate molars (N=115), and squamate vertebrae (N=254) under aligned, mildly perturbed, and severely perturbed initialization. We evaluated these formulations in terms of landmark accuracy, failure-tail behavior, runtime, and preservation of morphometric structure. Under severe perturbation, we also evaluated an integrated recovery procedure that optimizes the starting template, recovers global pose, and then applies SSM-CPD with refinement. Severe misalignment produced widespread failures in pose-matched conventional CPD, whereas SSM regularization reduced median error from 13 - 21% to approximately 1 - 3% of the target bounding-box diagonal. This improvement was highly consistent for skulls and molars. In the morphologically complex vertebral dataset, however, low median error concealed a substantial minority of catastrophic failures (28% of vertebrae above a 6% error threshold), which integrated recovery reduced to 4%. Across datasets, automated landmarks preserved the major axes and pairwise relationships of morphometric shape space, and local CPD refinement improved covariance preservation. Embedding learned biological variation directly into registration broadens the range of starting conditions and anatomies for which automated landmark transfer remains reliable. Our results also show that failure tails and downstream morphometric fidelity must be assessed alongside median accuracy when automated landmarking methods are evaluated.

The origin of multicellular animals from unicellular ancestors remains a major evolutionary question. To address this transition, a reconstruction of the nature of the last unicellular ancestor of animals (LUAA) is key, and requires broad sampling of extant unicellular relatives, which together with animals form the clade Holozoa. Environmental surveys suggest that unicellular Holozoa diversity extends beyond the five groups with characterized members (Choanoflagellatea, Ichthyosporea, Filasterea, Corallochytrea, and Atreyea). However, the extent of the environmental diversity within and outside these groups remains unclear. To fill this gap, we here systematically re-evaluated Holozoa diversity using environmental 18S rDNA sequences compiled from updated datasets, including long-read metabarcoding surveys. We report nine uncharacterized lineages inside Holozoa (seven newly identified here), provisionally placed outside the described groups. We also uncovered a previously unreported diverse lineage deeper inside Opisthokonta, separate from Holozoa and Fungi. In addition, we identified 21 uncharacterized lineages within Choanoflagellatea and Ichthyosporea (eleven newly reported here). We used short-read metabarcoding datasets to characterize the environmental and geographical distributions of these lineages. Many are globally distributed and some can reach relative abundances comparable to those of the described groups, while others are restricted to specific environments. Our findings reveal substantial uncharacterized diversity missing from the cultured taxa used in comparative studies of animal origins. By placing this diversity in a phylogenetic and environmental context, our study will guide the targeted isolation efforts needed to broaden the taxon sampling available to reconstruct LUAA and unravel the unicellular-to-multicellular transition that gave rise to animals.

Abstract 1. Biomes are continental-scale vegetation units widely used in ecological and evolutionary research, often together with large-scale species-distribution data. Many conceptually distinct biome schemes exist, differing in the underlying definition, delineation methodology and spatial resolution, yet most studies using biomes default to a few readily available schemes rather than choosing the one best suited to the research question. 2. We present biomes, an R package distributing spatially explicit raster layers of 31 published global biome schemes in a harmonised format at a 10 x 10 km resolution, together with metadata and a concise description. To facilitate their use, biomes provides a user-friendly classification workflow to classify occurrence records into specific biomes and a data-driven ranking algorithm to help researchers to choose the most suitable biome scheme for their data and research question in a transparent, reproducible way. 3. The biomes workflow can be applied to any dataset of georeferenced terrestrial occurrences. It (1) accepts either user-provided occurrence records or records downloaded by the package, (2) ranks the schemes according to their fit to the occurrence data, (3) assigns the records to specific biomes, and (4) summarises the resulting taxon-biome patterns. We demonstrate this workflow using 17,030 occurrence records of 185 species of Bombacoideae (Malvaceae). The analysis identified the vegetation-based biome scheme of Ramankutty and Foley (1999) as the best fit and recovered the forest and savanna biome signal known in the group from the literature, with tropical evergreen woodland being the biome with the largest share of records (46.1%) and highest number of species (162 of 185 species; 87.6%). 4. biomes provides user-friendly access to technically standardised and spatially explicit layers of 31 biome schemes, and thereby enables the transparent, data-driven choice of biome schemes tailored to a specific research data and questions, improving the reproducibility and the specificity of biome use across ecological and evolutionary research.

Phylogenetic trees inferred from different genomic regions, or from organellar and nuclear genomes, frequently disagree owing to incomplete lineage sorting, hybridization, introgression or horizontal gene transfer. Such discordance is informative, yet the standard tool for comparing trees visually, the tanglegram, is restricted to two trees and is known to misrepresent their similarity. Here we introduce phylogenetic parallelograms, drawings in which any number of rooted trees are embedded into a common scaffold, a rooted phylogenetic network that displays all of them, and are drawn in parallel so that shared branches align while conflicting branches visibly diverge. On synthetic tree pairs, the complexity of a parallelogram tracks the topological distance between the trees, whereas the complexity of an optimized tanglegram does not. We apply the approach to genome-wide introgression in the Anopheles gambiae species complex and to organellar discordance in cats, grasses and Fagaceae. PhyloParallelograms is an open-source interactive application.

Abstract Habitat configuration influences population connectivity and, consequently, genetic structure. River networks provide heterogeneous, hierarchically arranged environments that shift drastically from upstream to downstream. We compared congeneric species across Ephemeroptera, Plecoptera, and Trichoptera, emphasizing longitudinal replacement (upstream to downstream) and the rarely studied wet rock (hygropetric) habitats. We surveyed six rivers on the Muroto Peninsula, Japan, qualitatively sampling aquatic insects at 63 sites. Cluster analysis based on the environmental data classified surveyed sites into four clusters. We analyzed a total of 10 species from four genera that exhibit longitudinal replacement patterns within genera. Genetic analyses based on the mitochondrial cytochrome c oxidase subunit I region revealed that upstream species showed higher genetic diversity than downstream species. In contrast, species adapted to hygropetric exhibited the lowest genetic differentiation among all habitat types. A novel contribution of this study is the inclusion of hygropetric species. The surprisingly low differentiation in hygropetric species suggests high connectivity, similar to lentic species. By comparing congeneric taxa across orders within environmentally similar rivers, we reduce phylogenetic and environmental confounds, strengthening inference that habitat configuration and dispersal traits jointly shape genetic structure. These findings provide a new perspective on riverine spatial ecology and underscore the importance of microhabitat-aware comparisons for evolutionary inference.

The palaeogeographic history of the Neotropics has inspired multiple diversification models, emphasizing either the Andes or Amazonia as primary drivers of biodiversity. However, the extensive Pebas wetland system that covered western Amazonia during the Miocene has gained attention as a potential barrier to dispersal and diversification. This system may have contributed to the modern diversity imbalance between the Andes and Amazonia, known as the Gentry pattern. Most evidence for this hypothesis comes from isolated case studies and fossils. Here, we conduct a meta-analysis including 1,361 species of Neotropical butterflies to evaluate the role of the Pebas wetland on the dynamics of dispersal and diversification. We found that Andean- and Amazonian-centred clades show complementary but temporarily distinct biogeographic patterns. High Andean diversity reflects an Andean origin, with early occupancy and diversification occurring within the Andes, followed by post-Pebas dispersal into Amazonia. By contrast, Amazonian-centred clades, reflect an Amazonian origin, with an early diversification in Amazonia and subsequent post-Pebas colonization of the Andes. While several clades show little evidence of Pebas influence, our study suggests that the Pebas wetland did not uniformly restrict diversification, but rather selectively constrained dispersal between the Andes and Amazonia, preventing simultaneous diversification in both regions and thereby creating a diversity imbalance that characterizes the modern Neotropic.

The Fusarium oxysporum species complex (FOSC) comprises numerous plant pathogens infecting crops such as tomato, lettuce, melon, and banana. Despite its economic and ecological importance and decades of research, the species names linked to different phylogenetic lineages within this species complex remain largely controversial. Earlier taxonomic frameworks relied on a limited number of marker genes, variable taxon sampling and inconsistent species definitions, leading to a fragmented and unstable taxonomy. To address these challenges, we here analysed nearly 2,500 conserved marker genes from 660 genomes, including genome assemblies from 22 ex-type strains that were sequenced using long-read sequencing technology, representing nearly half of the 51 currently described species within FOSC. Maximum likelihood- and coalescence-based analyses of the full phylogenomic data matrix as well as subsets yielded a robust phylogeny of FOSC, which was furthermore corroborated by whole genome comparisons. Many previously described species are redundant, while two novel lineages remain undescribed. Direct comparisons to traditional marker gene-based approaches demonstrate their inadequacy for species delimitation in FOSC. We show that our phylogenomic framework can guide the detection of novel marker genes able to distinguish the major lineages within FOSC, which will be essential for taxon delineation and classification in future studies. Thus, our phylogenomic analyses provide a stable cornerstone for revising FOSC taxonomy and offers a critical resource for the broader research community.

The fieldwork was carried out between December 2023 and June 2025, with the aim, among others, of confirming the possible presence of the North African Gerbil (Gerbillus campestris) in Melilla. To this end, 56 camera trap stations were installed, each equipped with a camera trap. Of the total number of camera trap stations installed, the presence of Gerbillus campestris was detected in six of them, representing a capture success rate of 10.71%. Following the detection of the species, a specimen was captured using cage traps. For identification purposes, samples of epithelial tissue from the ear and faeces were collected. These samples were sent to the Institute of Vertebrate Biology of the Czech Academy of Sciences, Brno, Czech Republic, for analysis. There, DNA was extracted from the samples using the Innu PREP DNA Mini Kit 2.0 (IST Innuscreen GmbH, Germany) according to the manufacturer's protocol. The results obtained confirm, for the first time through records of the species, the presence of Gerbillus campestris in Melilla and, therefore, in Spain and Europe.

The Siamese fighting fish (Betta splendens) is well-known for its high levels of aggression and complex, stereotyped displays, making it an ideal model for testing deep learning-based tools for high-throughput, unbiased quantification of aggressive behaviour. A deep-learning object recognition model (BettaAI) was developed to detect the species' aggressive displays, including threat (opercula opening and fin distension), attack (bites) behaviours, as well as surface air breathing, a behaviour correlated with the frequency and intensity of aggression. Videos of males confronting their mirror image were acquired simultaneously with synchronised top and side cameras. Two separate models were trained, with some behaviours identified from top-view frames and others from side-view frames; outputs from both models were combined and post-processed for final quantification. BettaAI performed with high accuracy: F1 scores were 0.99 for the top-view model and 0.93 for the side-view model. Average Precision (AP) at 0.5 IoU exceeded 0.98 for all categories except bites, a rare behaviour in the dataset, which reached 0.865. Comparisons with a manually curated "ground-truth" dataset from videos of 10 males showed strong concordance between automated quantification and manual annotations by expert observers across all behaviours. A case study on 461 males reacting either to a mirror or to a video playback of a conspecific demonstrated the high-throughput and unbiased advantages of BettaAI, with threat displays being more frequent toward the mirror image and attack rates not differing between treatments. This work underscores the potential of BettaAI for automated, bias-free quantification of aggressive displays in B. splendens.

Hydractiniidae is a diverse family of colonial hydrozoans exhibiting substantial variation in colony organization, reproduction and life-cycle strategies, including species with and without a medusa stage. Several hydractiniids are also established model organisms, yet species boundaries and distributions remain incompletely resolved, limiting comparisons of reproductive and developmental traits across taxa. Here, we combine morphological and molecular data to reassess hydractiniid diversity, species identities, and life-cycle variation in material from the North Atlantic and Arctic Ocean. We integrate detailed observations of colonies and reproductive structures with DNA barcoding and mitochondrial genomes to evaluate species boundaries and provide molecular support for taxonomic identifications. Our results confirm 10 species in the study area, occurring from shallow waters to 2600 m depth and from coastal fjords to deep-sea hydrothermal vents. We further show that some historical species identifications and distribution records require reassessment. We establish a new genus for Arctic hydractiniid species (Bonneviactinia gen. nov.) and reassign two species to Schuchertinia. Mapping reproductive strategies onto the phylogeny reveals a broad pattern of increasing retention of medusa characters, from fixed sporosacs through reduced and fixed medusoids and non-liberated eumedusoids to fully developed, free-swimming feeding medusae, suggesting a progressive evolutionary trajectory in life-cycle complexity across Hydractiniidae. Furthermore, we clarify the identity of Podocoryna carnea and, through rearing experiments from polyp stage to medusae, characterize its morphological variation and developmental stages. By linking morphologically documented specimens and vouchers to genetic data, we establish reference resources that will facilitate future taxonomic, phylogenetic and biodiversity studies, including molecular surveys of marine communities. Together, these results demonstrate the value of integrative taxonomy for resolving hydractiniid diversity and provide a comparative foundation for investigating the evolution of reproductive strategies and life-cycle complexity in Medusozoa.

1. Occurrence records underpin biodiversity analyses, but their coordinates are often erroneous or missing. Misplaced coordinates distort environmental predictors and range estimates used in conservation assessment. Existing tools flag suspect records or georeference locality text, but rarely correct coordinates, provide calibrated uncertainty, or quantify downstream consequences. 2. We present GeoMend, a deterministic, rule-based method that flags and corrects suspect coordinates and georeferences records lacking coordinates from locality text, assigning each a precision tier or uncertainty radius. It uses no machine learning or artificial intelligence. We evaluate it against expert coordinates for 16,913 records and 736 species on Hispaniola, compare georeferencers using 5,000 localities from 72 countries, and measure changes in environmental predictors and the extent of occurrence. 3. On held-out Hispaniola records, GeoMend flags 92% of coordinates more than 5 km from expert truth, versus 27% for CoordinateCleaner. Among the accepted corrections with paired coordinates, the median error falls from 2.19 to 0.82 km, and 68% of the complete held-out pairs improve. On records placed by both GeoMend and GEOLocate that also have original coordinates, the median error is 0.67 km for GeoMend versus 10.69 km for GEOLocate. For records lacking coordinates, evaluated uncertainty radii contain expert truth about 92% of the time in well-supported gazetteer classes. Correction reduces the mean standardized environmental distance from expert truth by 38% and moves the extent of occurrence toward expert truth by a median of 8% (mean of 79%). Corrections and newly georeferenced records give 17 Hispaniola species their first range estimate. 4. On the global benchmark, GeoMend is more accurate than GPT-4o for two-thirds of the records placed by all four comparators, with median errors of 4.4 and 10.8 km, respectively. Applied to Global Biodiversity Information Facility (GBIF) records lacking coordinates, it places records from 82,180 previously unmapped species across 233 countries, ranging from precise locality interpretations to administrative centroids. Automated placement makes the global backlog more approachable, while explicit tiers distinguish uses supported by precise and coarse coordinates. Further georeferencing requires regional expertise to combine automated placement with local knowledge.

The twisted-wing parasites (order Strepsiptera) are well-known for their unusual life histories as obligate endoparasites of a wide variety of other insect orders. Here, we present the first high-quality genome assembly of a leafhopper-parasitizing strepsipteran species. We sequenced the genome of Elenchus koebelei using PacBio HiFi technology, yielding a primary assembly of 63.7Mb in 96 contigs, with an N50 of 1.60Mb and a BUSCO completeness score of 82.3%. In total, we recovered 8,167 protein-coding gene annotations, and 34.2% of the genome annotated as repeats. For the first time, we compare the genome content of three strepsipteran species and find that they are remarkably consistent in broad structure, with approximately 600 insect universal single-copy orthologs that appear to be completely missing from the order. These similarities imply a conserved genomic blueprint for strepsipteran parasitic strategy and/or differences that must be investigated at the post-translational level.

In this paper, we introduce Nagini3D, a 3D biological image segmentation method that hybridizes two high-performing yet highly complementary segmentation paradigms: deep learning-based approaches and snake-based methods. Our approach leverages both the high efficiency of data-driven techniques and the convenient shape representation provided by active surfaces, which proves highly practical for incorporating a priori knowledge as well as for the post-analysis of segmented objects. By considering a complete 3D volume, our method ensures segmentation consistency across all three dimensions, unlike many approaches that handle volumes in a slice-by-slice manner. Additionally, it theoretically enables the representation of any object with spherical topology, overcoming the limitations of common 3D methods that rely on object convexity assumptions. Furthermore, the continuous representation of objects allows for a resolution-independent description, facilitating the extraction of various geometric features of interest (e.g., local curvature, principal deformations). We demonstrate both the quality of the segmentation produced by our method and the richness of the accessible geometric features through an extensive set of experiments conducted on multiple real and simulated datasets, including objects with challenging geometries.

Automated monitoring camera systems offer an efficient approach for quantifying pollinator biodiversity and plant-pollinator interactions, but financial and logistical constraints limit the number of flowering plant species that can be monitored. Using a large European database of plant-pollinator networks, we evaluated whether plant subsampling can capture key metrics of interest (i.e. pollinator richness and network structure). We compared abundance-based, flower-shape-informed, phylogenetically informed and random plant sampling. Abundance-based sampling consistently outperformed random sampling, while adding flower shape provided little additional benefit and phylogenetic selection performed similarly to random sampling. Overall, monitoring 12-15 flowering species was sufficient to characterize key network properties across varying community sizes. Performance of abundance-based plant sampling declined in species-rich communities and when rare but highly attractive plants were present in the community, while a specific subset of pollinator species was consistently missed even by the best sampling strategy. Based on their relative contribution to interactions within each network, only 12.6% of pollinator species accounted for 95% of interactions, suggesting that AI classifiers could prioritize a relatively small subset of species. Our results provide practical guidance for designing efficient camera-based pollinator monitoring schemes at large spatial scales.

1. Learned sequence embeddings are increasingly proposed for fungal ITS classification and novelty detection, but are usually evaluated against weaker versions of themselves or default-configured alignment, by AUROC alone, and with queries treated as independent. We asked whether a purpose-trained encoder outperforms percent identity when both score identical queries under a firewalled open-world design, and which evaluation choices decide the answer. 2. From the UNITE release of 19 February 2025 we built ITS-core, ITS1 and ITS2 views and a genus-separated split with sealed calibration and test partitions. A convolutional encoder trained with genus-proxy, cross-view, hierarchical and episodic objectives was frozen and hash-sealed before test data were opened. We compared it with exhaustive, coverage-filtered VSEARCH identity on the same queries and references using genus-clustered paired intervals, and logged every post-opening correction before computing corrected metrics. 3. Three evaluation choices changed the comparison. Default VSEARCH heuristics returned a lower-identity hit than exhaustive search for 48.3% of benchmark queries; without a coverage filter, identity placed only 64.9% of novel ITS-core queries in the correct class, against 98.7% with it, because the conserved 5.8S let partial alignments win; and the encoder's embeddings depended on inference batching. Once corrected, identity exceeded the encoder in known-genus accuracy and novel-family placement at every view (family 80.0% to 84.6% against 53.8% to 64.8%) and detected more novel genera at a lower false-novelty rate. The encoder's novelty AUROC was within 0.023 of identity's, and no genus-clustered interval excluded zero. Identity's own development-to-test gap exceeded the encoder's, so that gap reflects partition composition rather than selection. Of the genera held out by a historical benchmark, 85% had been present in training; on a leakage-safe version, identity's AUROC advantage was 0.065, with a genus-clustered interval excluding zero. 4. Correctly configured alignment matched or exceeded the learned embedding throughout. The decisive results came from the evaluation, not the representation: baseline configuration, batch invariance, genus-level uncertainty, a parameter-free control for selection and a leakage audit each changed a conclusion, and each is inexpensive to apply to any learned barcode method.