Identification of Species by Combining Molecular and Morphological Data Using Convolutional Neural Networks.

Integrative taxonomy is central to modern taxonomy and systematic biology, including behavior, niche preference, distribution, morphological analysis, and DNA barcoding. However, decades of use demonstrate that these methods can face challenges when used in isolation, for instance, potential misidentifications due to phenotypic plasticity for morphological methods, and incorrect identifications because of introgression, incomplete lineage sorting, and horizontal gene transfer for DNA barcoding. Although researchers have advocated the use of integrative taxonomy, few detailed algorithms have been proposed. Here, we develop a convolutional neural network method (morphology-molecule network [MMNet]) that integrates morphological and molecular data for species identification. The newly proposed method (MMNet) worked better than four currently available alternative methods when tested with 10 independent data sets representing varying genetic diversity from different taxa. High accuracies were achieved for all groups, including beetles (98.1% of 123 species), butterflies (98.8% of 24 species), fishes (96.3% of 214 species), and moths (96.4% of 150 total species). Further, MMNet demonstrated a high degree of accuracy ($>$98%) in four data sets including closely related species from the same genus. The average accuracy of two modest subgenomic (single nucleotide polymorphism) data sets, comprising eight putative subspecies respectively, is 90%. Additional tests show that the success rate of species identification under this method most strongly depends on the amount of training data, and is robust to sequence length and image size. Analyses on the contribution of different data types (image vs. gene) indicate that both morphological and genetic data are important to the model, and that genetic data contribute slightly more. The approaches developed here serve as a foundation for the future integration of multimodal information for integrative taxonomy, such as image, audio, video, 3D scanning, and biosensor data, to characterize organisms more comprehensively as a basis for improved investigation, monitoring, and conservation of biodiversity. [Convolutional neural network; deep learning; integrative taxonomy; single nucleotide polymorphism; species identification.].

The impact of heavy metal pollution on wild bee communities in smallholder farmlands.

Wild bees provide important pollination services, but they face numerous stressors that threaten them and their ecosystem services. Wild bees can be exposed to heavy metal pollution through the consumption of nectar, pollen, and water, which might cause bee decline. While some studies have measured heavy metal concentrations in honeybees, few studies have monitored heavy metal concentrations in wild bees or explored their potential effects on wild bee communities. To investigate the impact of heavy metal pollution on wild bee communities, heavy metal concentrations, including vanadium (V), chromium (Cr), nickel (Ni), cadmium (Cd), Zinc (Zn) and lead (Pb) in multiple wild bee species were measured. Multiple wild bee species, including: Xylocopa tranquabaroroum, Eucera floralia, Apis cerana, and small bee mixtures (representing multiple small wild bee species) were sampled from 18 sites in Quzhou, Zhejiang Province, China. The findings demonstrated that there were significant differences in heavy metal concentrations among different bee species. The concentrations of V, Zn, Cd, and Pb in X. tranquabaroroum, the largest bee species in this study, were lower than that in the other three sample groups. Furthermore, there were significant negative correlations between heavy metal pollution and wild bee diversity and species richness, but not with abundance. Particularly, there was no significant relationship between heavy metal pollution and the abundance of small bees. Given these worrying findings, monitoring multiple heavy metals in wild bees should be conducted for protecting wild bee diversity and securing their pollination services.

Comparison of two criteria on the essential number calculation of Andrena camellia.

Andrena camellia Wu is one of the primary pollinators of Camellia oleifera A. in China. In this paper, the essential number of individuals for efficient pollination by this species was calculated via two criteria, based on various indicators including counts of pollen grains in provisions, from single visits, and from single foraging trips overall; single flower visit duration; single flight period duration; number of eggs laid by a single female over their lifetime; and the average number of flowers per plant. Based on the number of pollen grains collected per flower visit, the essential number of females necessary is 2107 in a 1-ha Camellia oleifera garden with 1800 plants, while only 1998 female individuals are essentially needed when estimated based on the mean number of pollen grains collected in a single flight period. We argue that the essential number estimated by the former method is more reasonable and accurate for practical applications.

Comparison of large-scale citizen science data and long-term study data for phenology modeling.

Large-scale observational data from citizen science efforts are becoming increasingly common in ecology, and researchers often choose between these and data from intensive local-scale studies for their analyses. This choice has potential trade-offs related to spatial scale, observer variance, and interannual variability. Here we explored this issue with phenology by comparing models built using data from the large-scale, citizen science USA National Phenology Network (USA-NPN) effort with models built using data from more intensive studies at Long Term Ecological Research (LTER) sites. We built statistical and process based phenology models for species common to each data set. From these models, we compared parameter estimates, estimates of phenological events, and out-of-sample errors between models derived from both USA-NPN and LTER data. We found that model parameter estimates for the same species were most similar between the two data sets when using simple models, but parameter estimates varied widely as model complexity increased. Despite this, estimates for the date of phenological events and out-of-sample errors were similar, regardless of the model chosen. Predictions for USA-NPN data had the lowest error when using models built from the USA-NPN data, while LTER predictions were best made using LTER-derived models, confirming that models perform best when applied at the same scale they were built. This difference in the cross-scale model comparison is likely due to variation in phenological requirements within species. Models using the USA-NPN data set can integrate parameters over a large spatial scale while those using an LTER data set can only estimate parameters for a single location. Accordingly, the choice of data set depends on the research question. Inferences about species-specific phenological requirements are best made with LTER data, and if USA-NPN or similar data are all that is available, then analyses should be limited to simple models. Large-scale predictive modeling is best done with the larger-scale USA-NPN data, which has high spatial representation and a large regional species pool. LTER data sets, on the other hand, have high site fidelity and thus characterize inter-annual variability extremely well. Future research aimed at forecasting phenology events for particular species over larger scales should develop models that integrate the strengths of both data sets.

Lessons from assembling UCEs: A comparison of common methods and the case of Clavinomia (Halictidae).

Sequence data assembly is a foundational step in high-throughput sequencing, with untold consequences for downstream analyses. Despite this, few studies have interrogated the many methods for assembling phylogenomic UCE data for their comparative efficacy, or for how outputs may be impacted. We study this by comparing the most commonly used assembly methods for UCEs in the under-studied bee lineage Nomiinae and a representative sampling of relatives. Data for 63 UCE-only and 75 mixed taxa were assembled with five methods, including ABySS, HybPiper, SPAdes, Trinity and Velvet, and then benchmarked for their relative performance in terms of locus capture parameters and phylogenetic reconstruction. Unexpectedly, Trinity and Velvet trailed the other methods in terms of locus capture and DNA matrix density, whereas SPAdes performed favourably in most assessed metrics. In comparison with SPAdes, the guided-assembly approach HybPiper generally recovered the highest quality loci but in lower numbers. Based on our results, we formally move Clavinomia to Dieunomiini and render Epinomia once more a subgenus of Dieunomia. We strongly advise that future studies more closely examine the influence of assembly approach on their results, or, minimally, use better-performing assembly methods such as SPAdes or HybPiper. In this way, we can move forward with phylogenomic studies in a more standardized, comparable manner.

Designing and Evaluating a Portable UV-LED Vane Trap to Expedite Arthropod Biodiversity Discovery.

A novel design of a portable funnel light trap (PFLT) was presented for collecting insects in ecological studies. The trap consists of a compact plastic box equipped with a light source and power source, along with two plastic polypropylene interception vanes. The PFLT costs 18.3 USD per unit and weighs approximately 300 g. A maximum of six PFLT units can be packed in one medium-sized backpack (32 cm × 45 cm × 15 cm, 20 L), making it easier to set up multiple units in remote areas wherein biodiversity research is needed. The low cost and weight of the trap also allows for large-scale deployment. The design is customizable and can be easily manufactured to fit various research needs. To validate the PFLT's efficacy in collecting insects across different habitat types, a series of field experiments were conducted in South Korea and Laos, where 37 trials were carried out. The PFLT successfully collected 7497 insects without experiencing battery issues or damage by rain or wind. Insect compositions and abundances differed across the three sampled habitat types: forests, grasslands, and watersides. This new FLT trap is an important tool for studying and protecting insect biodiversity, particularly in areas wherein conventional light traps cannot be deployed.

Global patterns and drivers of buzzing bees and poricidal plants.

Foraging behavior frequently plays a major role in driving the geographic distribution of animals. Buzzing to extract protein-rich pollen from flowers is a key foraging behavior used by bee species across at least 83 genera (these genera comprise ∼58% of all bee species). Although buzzing is widely recognized to affect the ecology and evolution of bees and flowering plants (e.g., buzz-pollinated flowers), global patterns and drivers of buzzing bee biogeography remain unexplored. Here, we investigate the global species distribution patterns within each bee family and how patterns and drivers differ with respect to buzzing bee species. We found that both distributional patterns and drivers of richness typically differed for buzzing species compared with hotspots for all bee species and when grouped by family. A major predictor of the distribution, but not species richness overall for buzzing members of four of the five major bee families included in analyses (Andrenidae, Halictidae, Colletidae, and to a lesser extent, Apidae), was the richness of poricidal flowering plant species, which depend on buzzing bees for pollination. Because poricidal plant richness was highest in areas with low wind and high aridity, we discuss how global hotspots of buzzing bee biodiversity are likely influenced by both biogeographic factors and plant host availability. Although we explored global patterns with state-level data, higher-resolution work is needed to explore local-level drivers of patterns. From a global perspective, buzz-pollinated plants clearly play a greater role in the ecology and evolution of buzzing bees than previously predicted.

Floral resource partitioning of coexisting bumble bees: Distinguishing species-, colony-, and individual-level effects.

Resource partitioning is considered a key factor in alleviating competitive interactions, enabling coexistence among consumer species. However, most studies have focused on resource partitioning between species, ignoring the potentially critical role of intraspecific variation in resource use. We investigated floral resource partitioning across species, colonies, and individuals in a species-rich bumblebee community in the diversification center of bumblebees. We used a total of 10,598 bumblebees belonging to 13 species across 5 years in the Hengduan Mountains of southwest China. First, we evaluated the influence of a comprehensive set of floral traits, including both those related to attractiveness (flower color and shape) and rewards (pollen, sugar ratio, nectar volume, sugar concentration, and amino acid content) on resource partitioning at the species level in bumblebee-plant networks. Then, we explored intraspecific resource partitioning on the colony and individual levels. Our results suggest that bumblebee species differ substantially in their use of the available floral resources, and that this mainly depends on flower attractiveness (floral color and shape). Interestingly, we also detected floral resource partitioning at the colony level within all commonest bumblebee species evaluated. In general, floral resource partitioning between bumblebee individuals decreased with species- and individual-level variation in body size (intertegular span). These results suggest that bumblebee species may coexist via the flexibility in their preferences for specific floral traits, which filters up to support the co-occurrence of high numbers of species and individuals in this global hotspot of species richness.

An integrative bioinformatics pipeline shows that honeybee-associated microbiomes are driven primarily by pollen composition.

The microbiomes associated with bee nests influence colony health through various mechanisms, although it is not yet clear how honeybee congeners differ in microbiome assembly processes, in particular the degrees to which floral visitations and the environment contribute to different aspects of diversity. We used DNA metabarcoding to sequence bacterial 16S rRNA from honey and stored pollen from nests of 4 honeybee species (Apis cerana, A. dorsata, A. florea, and A. laboriosa) sampled throughout Yunnan, China, a global biodiversity hotspot. We developed a computational pipeline integrating multiple databases for quantifying key facets of diversity, including compositional, taxonomic, phylogenetic, and functional ones. Further, we assessed candidate drivers of observed microbiome dissimilarity, particularly differences in floral visitations, habitat disturbance, and other key environmental variables. Analyses revealed that microbiome alpha diversity was broadly equivalent across the study sites and between bee species, apart from functional diversity which was very low in nests of the reclusive A. laboriosa. Turnover in microbiome composition across Yunnan was driven predominantly by pollen composition. Human disturbance negatively impacted both compositional and phylogenetic alpha diversity of nest microbiomes, but did not correlate with microbial turnover. We herein make progress in understanding microbiome diversity associated with key pollinators in a biodiversity hotspot, and provide a model for the use of a comprehensive informatics framework in assessing pattern and drivers of diversity, which enables the inclusion of explanatory variables both subtly and fundamentally different and enables elucidation of emergent or unexpected drivers.

Launching insectphylo.org; a new hub facilitating construction and use of synthesis molecular phylogenies of insects.

The Holy Grail of an Insect Tree of Life can only be 'discovered' through extensive collaboration among taxon specialists, phylogeneticists and centralized frameworks such as Open Tree of Life, but insufficient effort from stakeholders has so far hampered this promising approach. The resultant unavailability of synthesis phylogenies is an unfortunate situation given the numerous practical usages of phylogenies in the near term and against the backdrop of the ongoing biodiversity crisis. To resolve this issue, we establish a new online hub that centralizes the collation of relevant phylogenetic data and provides the resultant synthesis molecular phylogenies. This is achieved through key developments in a proposed pipeline for the construction of a species-level insect phylogeny. The functionality of the framework is demonstrated through the construction of a highly supported, species-comprehensive phylogeny of Diptera, built from integrated omics data, COI DNA barcodes, and a compiled database of over 100 standardized, published Diptera phylogenies. Machine-readable forms of the phylogeny (and subsets thereof) are publicly available at insectphylo.org, a new public repository for species-comprehensive phylogenies for biological research.

The Whole Genome of Colletes collaris (Hymenoptera: Colletidae): An Important Step in Comparative Genomics of Cellophane Bees.

Gaining knowledge on bees is of the utmost importance due to the paramount role that they play in angiosperm pollination. Herein, we provide the first genome assembly of Colletes collaris, a pan-Eurasian cellophane bee. We sequenced 50.53 Gbp of long-read data plus 57.36 Gbp of short-read data in Oxford Nanopore Technologies and Illumina platforms, respectively. The genome assembly consisted of 374.75 Mbp distributed across 374 contigs, with L50 and N50 of 9 and 8.96 Mbp, respectively. We predicted the genome to comprise 20,399 protein-coding genes, 467,947 repeats, and 4,315 non-coding RNA genes. The transcriptome and mitochondrial genome of the species were also assembled. Gene family analysis with 15 insect species identified 14,417 families, 9,517 of them found in C. collaris. A dated phylogenomic analysis revealed high numbers of orthogroups experiencing rapid evolution within Colletes.

Phylogeny-based assignment of functional traits to DNA barcodes outperforms distance-based, in a comparison of approaches.

The full potential for using DNA barcodes for profiling functional trait diversity has yet to be determined in plants and animals; thus, we outline a general framework for quantifying functional trait diversity of insect community DNA and propose and assess the accuracy of three methods for achieving this. We built a novel dataset of traits and DNA barcodes for wild bees in China. An informatics framework was developed for phylogeny-based integration of these data and prediction of traits for any subject barcodes, which was compared with two distance-based methods. For Phylogenetic Assignment, we additionally conducted a species-level analysis of publically available bee trait data. Under the specimen-level dataset, the rate of trait assignment was negatively correlated with distance between the query and the nearest trait-known reference, for all methods. Phylogenetic Assignment was found to perform best under several criteria; particularly, it had the lowest false-positive rate (rarely returning a state prediction where success was unlikely; where the distance from query to the nearest reference was high). For a wider range of compiled traits, conservative life-history traits showed the highest rates of assignment; for example, sociality was predicted with confidence at 53%, parasitism at 44% and nest location at 33%. As outlined herein, automated trait assignment might be applied at scale to either barcodes or metabarcodes. With further compilation and databasing of DNA barcode and trait data, the rate and accuracy of trait assignment is expected to increase to the point of being a widely viable and informative approach.

The evolutionary history of bees in time and space.

Bees are the most significant pollinators of flowering plants. This partnership began ca. 120 million years ago, but the uncertainty of how and when bees spread across the planet has greatly obscured investigations of this key mutualism. We present a novel analysis of bee biogeography using extensive new genomic and fossil data to demonstrate that bees originated in Western Gondwana (Africa and South America). Bees likely originated in the Early Cretaceous, shortly before the breakup of Western Gondwana, and the early evolution of any major bee lineage is associated with either the South American or African land masses. Subsequently, bees colonized northern continents via a complex history of vicariance and dispersal. The notable early absences from large landmasses, particularly in Australia and India, have important implications for understanding the assembly of local floras and diverse modes of pollination. How bees spread around the world from their hypothesized Southern Hemisphere origin parallels the histories of numerous flowering plant clades, providing an essential step to studying the evolution of angiosperm pollination syndromes in space and time.

A globally synthesised and flagged bee occurrence dataset and cleaning workflow.

Species occurrence data are foundational for research, conservation, and science communication, but the limited availability and accessibility of reliable data represents a major obstacle, particularly for insects, which face mounting pressures. We present BeeBDC, a new R package, and a global bee occurrence dataset to address this issue. We combined >18.3 million bee occurrence records from multiple public repositories (GBIF, SCAN, iDigBio, USGS, ALA) and smaller datasets, then standardised, flagged, deduplicated, and cleaned the data using the reproducible BeeBDC R-workflow. Specifically, we harmonised species names (following established global taxonomy), country names, and collection dates and, we added record-level flags for a series of potential quality issues. These data are provided in two formats, "cleaned" and "flagged-but-uncleaned". The BeeBDC package with online documentation provides end users the ability to modify filtering parameters to address their research questions. By publishing reproducible R workflows and globally cleaned datasets, we can increase the accessibility and reliability of downstream analyses. This workflow can be implemented for other taxa to support research and conservation.

Searching the web builds fuller picture of arachnid trade.

Wildlife trade is a major driver of biodiversity loss, yet whilst the impacts of trade in some species are relatively well-known, some taxa, such as many invertebrates are often overlooked. Here we explore global patterns of trade in the arachnids, and detected 1,264 species from 66 families and 371 genera in trade. Trade in these groups exceeds millions of individuals, with 67% coming directly from the wild, and up to 99% of individuals in some genera. For popular taxa, such as tarantulas up to 50% are in trade, including 25% of species described since 2000. CITES only covers 30 (2%) of the species potentially traded. We mapped the percentage and number of species native to each country in trade. To enable sustainable trade, better data on species distributions and better conservation status assessments are needed. The disparity between trade data sources highlights the need to expand monitoring if impacts on wild populations are to be accurately gauged and the impacts of trade minimised.

Chromosome-Level Genome Assembly of Anthidium xuezhongi Niu & Zhu, 2020 (Hymenoptera: Apoidea: Megachilidae: Anthidiini).

Anthidiini, a large bee tribe characterized by light-colored maculations, represents nearly 1,000 pollinator species, but no genomes are yet available for this tribe. Here, we report a chromosome-level genome assembly of Anthidium xuezhongi collected from the Tibetan Plateau. Using PacBio long reads and Hi-C data, we assembled a genome of 189.14 Mb with 99.94% of the assembly located in 16 chromosomes. Our assembly contains 23 scaffolds, with the scaffold N50 length of 12.53 Mb, and BUSCO completeness of 98.70% (n = 1,367). We masked 25.98 Mb (13.74%) of the assembly as repetitive elements, identified 385 noncoding RNAs, and predicted 10,820 protein-coding genes (99.20% BUSCO completeness). Gene family evolution analyses identified 9,251 gene families, of which 31 gene families experienced rapid evolution. Interspecific chromosomal variation among A. xuezhongi, Bombus terrestris, and Apis mellifera showed strong chromosomal syntenic relationships. This high-quality genome assembly is a valuable resource for evolutionary and comparative genomic analyses of bees.

Testing the systematic status of Homalictus and Rostrohalictus with weakened cross-vein groups within Halictini (Hymenoptera: Halictidae) using low-coverage whole-genome sequencing.

The halictid genus Lasioglossum, as one of the most species-rich bee groups with persistently contentious subgeneric boundaries, is one of the most challenging bee groups from a systematic standpoint. An enduring question is the relationship of Lasioglossum and Homalictus, whether all halictine bees with weakened distal wing venation comprise one or multiple genera. Here, we analyzed the phylogenetic relationships among the subgroups within Lasioglossum s.l. based on thousands of single-copy orthologs and ultraconserved elements, which were extracted from 23 newly sequenced low-coverage whole genomes alongside a published genome (22 ingroups plus 2 outgroups). Both marker sets provided consistent results across maximum likelihood and coalescent-based species tree approaches. The phylogenetic and topology test results show that the Lasioglossum and Hemihalictus series are reciprocally monophyletic and Homalictus and Rostrohalictus are valid subgenera of Lasioglossum. Consequently, we lower Homalictus to subgenus status within Lasioglossum again, and we also raise Rostrohalictus to subgenus status from its prior synonymy with subgenus Hemihalictus. Lasioglossum przewalskyi is also transferred to the subgenus Hemihalictus. Ultimately, we redefine Lasioglossum to include all halictine bees with weakened distal wing venation.

Evidence for multiple introductions of an invasive wild bee species currently under rapid range expansion in Europe.

BACKGROUND: Invasive species are increasingly driving biodiversity decline, and knowledge of colonization dynamics, including both drivers and dispersal modes, are important to prevent future invasions. The bee species Megachile sculpturalis (Hymenoptera: Megachilidae), native to East-Asia, was first recognized in Southeast-France in 2008, and has since spread throughout much of Europe. The spread is very fast, and colonization may result from multiple fronts. RESULT: To track the history of this invasion, codominant markers were genotyped using Illumina sequencing and the invasion history and degree of connectivity between populations across the European invasion axis were investigated. Distinctive genetic clusters were detected with east-west differentiations in Middle-Europe. CONCLUSION: We hypothesize that the observed cluster formation resulted from multiple, independent introductions of the species to the European continent. This study draws a first picture of an early invasion stage of this wild bee and forms a foundation for further investigations, including studies of the species in their native Asian range and in the invaded range in North America.