Bridging the Gap between Field Experiments and Machine Learning: The EC H2020 B-GOOD Project as a Case Study towards Automated Predictive Health Monitoring of Honey Bee Colonies.

Honey bee colonies have great societal and economic importance. The main challenge that beekeepers face is keeping bee colonies healthy under ever-changing environmental conditions. In the past two decades, beekeepers that manage colonies of Western honey bees (Apis mellifera) have become increasingly concerned by the presence of parasites and pathogens affecting the bees, the reduction in pollen and nectar availability, and the colonies' exposure to pesticides, among others. Hence, beekeepers need to know the health condition of their colonies and how to keep them alive and thriving, which creates a need for a new holistic data collection method to harmonize the flow of information from various sources that can be linked at the colony level for different health determinants, such as bee colony, environmental, socioeconomic, and genetic statuses. For this purpose, we have developed and implemented the B-GOOD (Giving Beekeeping Guidance by computational-assisted Decision Making) project as a case study to categorize the colony's health condition and find a Health Status Index (HSI). Using a 3-tier setup guided by work plans and standardized protocols, we have collected data from inside the colonies (amount of brood, disease load, honey harvest, etc.) and from their environment (floral resource availability). Most of the project's data was automatically collected by the BEEP Base Sensor System. This continuous stream of data served as the basis to determine and validate an algorithm to calculate the HSI using machine learning. In this article, we share our insights on this holistic methodology and also highlight the importance of using a standardized data language to increase the compatibility between different current and future studies. We argue that the combined management of big data will be an essential building block in the development of targeted guidance for beekeepers and for the future of sustainable beekeeping.

Viral Co-Infections and Antiviral Immunity in Honey Bees.

Over the past few decades, honey bees have been facing an increasing number of stressors. Beyond individual stress factors, the synergies between them have been identified as a key factor in the observed increase in colony mortality. However, these interactions are numerous and complex and call for further research. Here, in line with our need for a systemic understanding of the threats that they pose to bee health, we review the interactions between honey bee viruses. As viruses are obligate parasites, the interactions between them not only depend on the viruses themselves but also on the immune responses of honey bees. Thus, we first summarise our current knowledge of the antiviral immunity of honey bees. We then review the interactions between specific pathogenic viruses and their interactions with their host. Finally, we draw hypotheses from the current literature and suggest directions for future research.

Virus Prevalence in Egg Samples Collected from Naturally Selected and Traditionally Managed Honey Bee Colonies across Europe.

Monitoring virus infections can be an important selection tool in honey bee breeding. A recent study pointed towards an association between the virus-free status of eggs and an increased virus resistance to deformed wing virus (DWV) at the colony level. In this study, eggs from both naturally surviving and traditionally managed colonies from across Europe were screened for the prevalence of different viruses. Screenings were performed using the phenotyping protocol of the 'suppressed in ovo virus infection' trait but with qPCR instead of end-point PCR and a primer set that covers all DWV genotypes. Of the 213 screened samples, 109 were infected with DWV, 54 were infected with black queen cell virus (BQCV), 3 were infected with the sacbrood virus, and 2 were infected with the acute bee paralyses virus. It was demonstrated that incidences of the vertical transmission of DWV were more frequent in naturally surviving than in traditionally managed colonies, although the virus loads in the eggs remained the same. When comparing virus infections with queen age, older queens showed significantly lower infection loads of DWV in both traditionally managed and naturally surviving colonies, as well as reduced DWV infection frequencies in traditionally managed colonies. We determined that the detection frequencies of DWV and BQCV in honey bee eggs were lower in samples obtained in the spring than in those collected in the summer, indicating that vertical transmission may be lower in spring. Together, these patterns in vertical transmission show that honey bee queens have the potential to reduce the degree of vertical transmission over time.

Epidemiology of the Microsporidium Nosema ceranae in Four Mediterranean Countries.

Nosema ceranae is a highly prevalent intracellular parasite of honey bees' midgut worldwide. This Microsporidium was monitored during a long-term study to evaluate the infection at apiary and intra-colony levels in six apiaries in four Mediterranean countries (France, Israel, Portugal, and Spain). Parameters on colony strength, honey production, beekeeping management, and climate were also recorded. Except for São Miguel (Azores, Portugal), all apiaries were positive for N. ceranae, with the lowest prevalence in mainland France and the highest intra-colony infection in Israel. A negative correlation between intra-colony infection and colony strength was observed in Spain and mainland Portugal. In these two apiaries, the queen replacement also influenced the infection levels. The highest colony losses occurred in mainland France and Spain, although they did not correlate with the Nosema infection levels, as parasitism was low in France and high in Spain. These results suggest that both the effects and the level of N. ceranae infection depends on location and beekeeping conditions. Further studies on host-parasite coevolution, and perhaps the interactions with other pathogens and the role of honey bee genetics, could assist in understanding the difference between nosemosis disease and infection, to develop appropriate strategies for its control.

Genetic diversification of an invasive honey bee ectoparasite across sympatric and allopatric host populations.

Invasive parasites are major threats to biodiversity. The honey bee ectoparasite, Varroa destructor, has shifted host and spread almost globally several decades ago. This pest is generally considered to be the main global threat to Western honey bees, Apis mellifera, although the damages it causes are not equivalent in all its new host's populations. Due to the high virulence of this parasite and the viruses it vectors, beekeepers generally rely on acaricide treatments to keep their colonies alive. However, some populations of A. mellifera can survive without anthropogenic mite control, through the expression of diverse resistance and tolerance traits. Such surviving colonies are currently found throughout the globe, with the biggest populations being found in Sub-Saharan Africa and Latin America. Recently, genetic differences between mite populations infesting surviving and treated A. mellifera colonies in Europe were found, suggesting that adaptations of honey bees drive mite evolution. Yet, the prevalence of such co-evolutionary adaptations in other invasive populations of V. destructor remain unknown. Using the previous data from Europe and novel genetic data from V. destructor populations in South America and Africa, we here investigated whether mites display signs of adaptations to different host populations of diverse origins and undergoing differing management. Our results show that, contrary to the differences previously documented in Europe, mites infesting treated and untreated honey bee populations in Africa and South America are genetically similar. However, strong levels of genetic differentiation were found when comparing mites across continents, suggesting ongoing allopatric speciation despite a recent spread from genetically homogenous lineages. This study provides novel insights into the co-evolution of V. destructor and A. mellifera, and confirms that these species are ideal to investigate coevolution in newly established host-parasite systems.

Honey bees and climate explain viral prevalence in wild bee communities on a continental scale.

Viruses are omnipresent, yet the knowledge on drivers of viral prevalence in wild host populations is often limited. Biotic factors, such as sympatric managed host species, as well as abiotic factors, such as climatic variables, are likely to impact viral prevalence. Managed and wild bees, which harbor several multi-host viruses with a mostly fecal-oral between-species transmission route, provide an excellent system with which to test for the impact of biotic and abiotic factors on viral prevalence in wild host populations. Here we show on a continental scale that the prevalence of three broad host viruses: the AKI-complex (Acute bee paralysis virus, Kashmir bee virus and Israeli acute paralysis virus), Deformed wing virus, and Slow bee paralysis virus in wild bee populations (bumble bees and solitary bees) is positively related to viral prevalence of sympatric honey bees as well as being impacted by climatic variables. The former highlights the need for good beekeeping practices, including Varroa destructor management to reduce honey bee viral infection and hive placement. Furthermore, we found that viral prevalence in wild bees is at its lowest at the extreme ends of both temperature and precipitation ranges. Under predicted climate change, the frequency of extremes in precipitation and temperature will continue to increase and may hence impact viral prevalence in wild bee communities.

Cold case: The disappearance of Egypt bee virus, a fourth distinct master strain of deformed wing virus linked to honeybee mortality in 1970's Egypt.

In 1977, a sample of diseased adult honeybees (Apis mellifera) from Egypt was found to contain large amounts of a previously unknown virus, Egypt bee virus, which was subsequently shown to be serologically related to deformed wing virus (DWV). By sequencing the original isolate, we demonstrate that Egypt bee virus is in fact a fourth unique, major variant of DWV (DWV-D): more closely related to DWV-C than to either DWV-A or DWV-B. DWV-A and DWV-B are the most common DWV variants worldwide due to their close relationship and transmission by Varroa destructor. However, we could not find any trace of DWV-D in several hundred RNA sequencing libraries from a worldwide selection of honeybee, varroa and bumblebee samples. This means that DWV-D has either become extinct, been replaced by other DWV variants better adapted to varroa-mediated transmission, or persists only in a narrow geographic or host range, isolated from common bee and beekeeping trade routes.

Variant A of the Deformed Wings Virus Alters the Olfactory Sensitivity and the Expression of Odorant Binding Proteins on Antennas of Apis mellifera.

Insects have a highly sensitive sense of smell, allowing them to perform complex behaviors, such as foraging and peer recognition. Their sense of smell is based on the recognition of ligands and is mainly coordinated by odorant-binding proteins (OBPs). In Apis mellifera, behavior can be affected by different pathogens, including deformed wing virus (DWV) and its variants. In particular, it has been shown that variant A of DWV (DWV-A) is capable of altering the ultra-cellular structure associated with olfactory activity. In this study was evaluated olfactory sensitivity and the expression of OBP genes in honey bees inoculated with DWV-A. Electroantennographic analyses (EAG) were carried out to determine the olfactory sensitivity to the essential oils Eucalyptus globulus and Mentha piperita. The expression of nine antenna-specific OBP genes and DWV-A load in inoculated bees was also quantified by qPCR. We observed an inverse relationship between viral load and olfactory sensitivity and the expression of some OBP proteins. Thus, high viral loads reduced olfactory sensitivity to essential oils and the gene expression of the OBP2, OBP5, OBP11, and OBP12 proteins on the antennas of middle- and forager-age bees. These results suggest that DWV-A could have negative effects on the processes of aroma perception by worker bees, affecting their performance in tasks carried out in and outside the colony.

A SNP assay for assessing diversity in immune genes in the honey bee (Apis mellifera L.).

With a growing number of parasites and pathogens experiencing large-scale range expansions, monitoring diversity in immune genes of host populations has never been so important because it can inform on the adaptive potential to resist the invaders. Population surveys of immune genes are becoming common in many organisms, yet they are missing in the honey bee (Apis mellifera L.), a key managed pollinator species that has been severely affected by biological invasions. To fill the gap, here we identified single nucleotide polymorphisms (SNPs) in a wide range of honey bee immune genes and developed a medium-density assay targeting a subset of these genes. Using a discovery panel of 123 whole-genomes, representing seven A. mellifera subspecies and three evolutionary lineages, 180 immune genes were scanned for SNPs in exons, introns (< 4 bp from exons), 3' and 5´UTR, and < 1 kb upstream of the transcription start site. After application of multiple filtering criteria and validation, the final medium-density assay combines 91 quality-proved functional SNPs marking 89 innate immune genes and these can be readily typed using the high-sample-throughput iPLEX MassARRAY system. This medium-density-SNP assay was applied to 156 samples from four countries and the admixture analysis clustered the samples according to their lineage and subspecies, suggesting that honey bee ancestry can be delineated from functional variation. In addition to allowing analysis of immunogenetic variation, this newly-developed SNP assay can be used for inferring genetic structure and admixture in the honey bee.

Intra-Colonial Viral Infections in Western Honey Bees (Apis Mellifera).

RNA viruses play a significant role in the current high losses of pollinators. Although many studies have focused on the epidemiology of western honey bee (Apis mellifera) viruses at the colony level, the dynamics of virus infection within colonies remains poorly explored. In this study, the two main variants of the ubiquitous honey bee virus DWV as well as three major honey bee viruses (SBV, ABPV and BQCV) were analyzed from Varroa-destructor-parasitized pupae. More precisely, RT-qPCR was used to quantify and compare virus genome copies across honey bee pupae at the individual and subfamily levels (i.e., patrilines, sharing the same mother queen but with different drones as fathers). Additionally, virus genome copies were compared in cells parasitized by reproducing and non-reproducing mite foundresses to assess the role of this vector. Only DWV was detected in the samples, and the two variants of this virus significantly differed when comparing the sampling period, colonies and patrilines. Moreover, DWV-A and DWV-B exhibited different infection patterns, reflecting contrasting dynamics. Altogether, these results provide new insight into honey bee diseases and stress the need for more studies about the mechanisms of intra-colonial disease variation in social insects.

Possible Spillover of Pathogens between Bee Communities Foraging on the Same Floral Resource.

Viruses are known to contribute to bee population decline. Possible spillover is suspected from the co-occurrence of viruses in wild bees and honey bees. In order to study the risk of virus transmission between wild and managed bee species sharing the same floral resource, we tried to maximize the possible cross-infections using Phacelia tanacetifolia, which is highly attractive to honey bees and a broad range of wild bee species. Virus prevalence was compared over two years in Southern France. A total of 1137 wild bees from 29 wild bee species (based on COI barcoding) and 920 honey bees (Apis mellifera) were checked for the seven most common honey bee RNA viruses. Halictid bees were the most abundant. Co-infections were frequent, and Sacbrood virus (SBV), Black queen cell virus (BQCV), Acute bee paralysis virus (ABPV) and Israeli acute paralysis virus (IAPV) were widespread in the hymenopteran pollinator community. Conversely, Deformed wing virus (DWV) was detected at low levels in wild bees, whereas it was highly prevalent in honey bees (78.3% of the samples). Both wild bee and honey bee virus isolates were sequenced to look for possible host-specificity or geographical structuring. ABPV phylogeny suggested a specific cluster for Eucera bees, while isolates of DWV from bumble bees (Bombus spp.) clustered together with honey bee isolates, suggesting a possible spillover.

Bee Viruses: Routes of Infection in Hymenoptera.

Numerous studies have recently reported on the discovery of bee viruses in different arthropod species and their possible transmission routes, vastly increasing our understanding of these viruses and their distribution. Here, we review the current literature on the recent advances in understanding the transmission of viruses, both on the presence of bee viruses in Apis and non-Apis bee species and on the discovery of previously unknown bee viruses. The natural transmission of bee viruses will be discussed among different bee species and other insects. Finally, the research potential of in vivo (host organisms) and in vitro (cell lines) serial passages of bee viruses is discussed, from the perspective of the host-virus landscape changes and potential transmission routes for emerging bee virus infections.

Interactions Between Thiamethoxam and Deformed Wing Virus Can Drastically Impair Flight Behavior of Honey Bees.

Exposure to multiple stress factors is believed to contribute to honey bee colony decline. However, little is known about how co-exposure to stress factors can alter the survival and behavior of free-living honey bees in colony conditions. We therefore studied the potential interaction between a neonicotinoid pesticide, thiamethoxam, and a highly prevalent honey bee pathogen, Deformed wing virus (DWV). For this purpose, tagged bees were exposed to DWV by feeding or injection, and/or to field-relevant doses of thiamethoxam, then left in colonies equipped with optical bee counters to monitor flight activity. DWV loads and the expression of immune genes were quantified. A reduction in vitellogenin expression level was observed in DWV-injected bees and was associated with precocious onset of foraging. Combined exposure to DWV and thiamethoxam did not result in higher DWV loads compared to bees only exposed to DWV, but induced precocious foraging, increased the risk of not returning to the hive after the first flight, and decreased survival when compared to single stress exposures. We therefore provided the first evidence for deleterious interactions between DWV and thiamethoxam in natural conditions.

Diversity and Global Distribution of Viruses of the Western Honey Bee, Apis mellifera.

In the past centuries, viruses have benefited from globalization to spread across the globe, infecting new host species and populations. A growing number of viruses have been documented in the western honey bee, Apis mellifera. Several of these contribute significantly to honey bee colony losses. This review synthetizes the knowledge of the diversity and distribution of honey-bee-infecting viruses, including recent data from high-throughput sequencing (HTS). After presenting the diversity of viruses and their corresponding symptoms, we surveyed the scientific literature for the prevalence of these pathogens across the globe. The geographical distribution shows that the most prevalent viruses (deformed wing virus, sacbrood virus, black queen cell virus and acute paralysis complex) are also the most widely distributed. We discuss the ecological drivers that influence the distribution of these pathogens in worldwide honey bee populations. Besides the natural transmission routes and the resulting temporal dynamics, global trade contributes to their dissemination. As recent evidence shows that these viruses are often multihost pathogens, their spread is a risk for both the beekeeping industry and the pollination services provided by managed and wild pollinators.

Viruses in the Invasive Hornet Vespa velutina.

The Asian yellow-legged hornet Vespa velutina nigrithorax, a major predator of honeybees, is spreading in Europe in part due to a lack of efficient control methods. In this study, as a first step to identify biological control agents, we characterized viral RNA sequences present in asymptomatic or symptomatic hornets. Among 19 detected viruses, the honey bee virus Deformed wing virus-B was predominant in all the samples, particularly in muscles from the symptomatic hornet, suggesting a putative cause of the deformed wing symptom. Interestingly, two new viruses closely related to Acyrthosiphon pisumvirus and Himetobi Pvirus and viruses typically associated with honey bees, Acute bee paralysis virus and Black queen cell virus, were detected in the brain and muscles, and may correspond to the circulation and possible replication forms of these viruses in the hornet. Aphid lethal paralysis virus, Bee Macula-like virus, and Moku virus, which are known to infect honey bees, were also identified in the gut virus metagenome of hornets. Therefore, our study underlined the urgent need to study the host range of these newly discovered viruses in hornets to determine whether they represent a new threat for honey bees or a hope for the biocontrol of V. velutina.

Influence of chronic exposure to thiamethoxam and chronic bee paralysis virus on winter honey bees.

Co-exposure to pesticides and viruses is likely to occur in honey bee colonies. Pesticides can be present in pollen, nectar, and persist in stored food (honey and bee bread), and viruses can be highly prevalent in honey bee colonies. Therefore, the present study describes the influence of chronic co-exposure to thiamethoxam and Chronic bee paralysis virus (CBPV) on bee survival, virus loads, expression level of immune and detoxication genes, and pesticide metabolism Experiments were performed on honey bees collected from a winter apiary with reduced viral contaminations. No synergistic effect of co-exposure was observed on bee survival, nor on the ability of bees to metabolise the pesticide into clothianidin. However, we found that co-exposure caused an increase in CBPV loads that reached the viral levels usually found in overt infections. The effect of co-exposure on CBPV replication was associated with down-regulation of vitellogenin and dorsal-1a gene transcription. Nevertheless, the observed effects might be different to those occurring in spring or summer bees, which are more likelyco-exposed to thiamethoxam and CBPV and exhibit a different physiology.

Hazard of a neonicotinoid insecticide on the homing flight of the honeybee depends on climatic conditions and Varroa infestation.

The paradigm for all toxicological bioassays in the risk assessment of pesticide registration reflects the principle that experimental conditions should be controlled to avoid any other factors that may affect the endpoint measures. As honeybee colonies can be frequently exposed to bio-aggressors in real conditions, often concomitantly with pesticides, co-exposure to pesticide/bio-aggressors is becoming a concern for regulatory authorities. We investigated the effects of the neonicotinoid insecticide thiamethoxam on the homing performances of foragers emerging from colonies differentiated by health status (infestation with Varroa destructor mites, microsporidian parasite Nosema spp. and Deformed Wing Virus). We designed a homing test that has been recently identified to fill a regulatory gap in the field evaluations of sublethal doses of pesticides before their registration. We also assessed the effect of temperature as an environmental factor. Our results showed that the Varroa mite exacerbates homing failure (HF) caused by the insecticide, whereas high temperatures reduce insecticide-induced HF. Through an analytical Effective Dose (ED) approach, predictive modeling results showed that, for instance, ED level of an uninfested colony, can be divided by 3.3 when the colony is infested by 5 Varroa mites per 100 bees and at a temperature of 24 °C. Our results suggest that the health status of honeybee colonies and climatic context should be targeted for a thorough risk assessment.

Temperature-driven changes in viral loads in the honey bee Apis mellifera.

Many of the physiological traits in insects are shaped by environmental temperatures, which can influence their interactions with pathogens. Therefore, quantifying the thermal responses of the host-pathogen system is crucial for better understanding and predicting their dynamics due to environmental changes. This is particularly important in honey bees, which are experiencing severe colony losses around the world, notably due to infection with the Deformed wing virus (DWV). To investigate the influence of temperature on the honey bee/DWV relationship we exposed adult bees to low or high temperatures and determined the effects on viral titers and bee survival. Emerging bees naturally infected with DWV were reared in vitro at different temperatures ranging from 15 °C to 37 °C. In addition, some bees reared at 37 °C were exposed daily to acute heat treatments (40 and 43 °C). High temperatures significantly decreased DWV titers close to the initial viral load at emergence but increased bee mortality. The lowest temperature resulted in higher mortality, but virus load was not significantly impacted. In conclusion, our results indicate that temperature could contribute to seasonal variations in viral loads but do not suggest temperature to be used as a tool to eliminate viruses, even given that high temperatures limit viral multiplication.

Nosema ceranae in Apis mellifera: a 12 years postdetection perspective.

Nosema ceranae is a hot topic in honey bee health as reflected by numerous papers published every year. This review presents an update of the knowledge generated in the last 12 years in the field of N. ceranae research, addressing the routes of transmission, population structure and genetic diversity. This includes description of how the infection modifies the honey bee's metabolism, the immune response and other vital functions. The effects on individual honey bees will have a direct impact on the colony by leading to losses in the adult's population. The absence of clear clinical signs could keep the infection unnoticed by the beekeeper for long periods. The influence of the environmental conditions, beekeeping practices, bee genetics and the interaction with pesticides and other pathogens will have a direct influence on the prognosis of the disease. This review is approached from the point of view of the Mediterranean countries where the professional beekeeping has a high representation and where this pathogen is reported as an important threat.

The discovery of Halictivirus resolves the Sinaivirus phylogeny.

By providing pollination services, bees are among the most important insects, both in ecological and economical terms. Combined next-generation and classical sequencing approaches were applied to discover and study new insect viruses potentially harmful to bees. A bioinformatics virus discovery pipeline was used on individual Illumina transcriptomes of 13 wild bees from three species from the genus Halictus and 30 ants from six species of the genera Messor and Aphaenogaster. This allowed the discovery and description of three sequences of a new virus termed Halictus scabiosae Adlikon virus (HsAV). Phylogenetic analyses of ORF1, RNA-dependent RNA-polymerase (RdRp) and capsid genes showed that HsAV is closely related to (+)ssRNA viruses of the unassigned Sinaivirus genus but distant enough to belong to a different new genus we called Halictivirus. In addition, our study of ant transcriptomes revealed the first four sinaivirus sequences from ants (Messor barbarus, M. capitatus and M. concolor). Maximum likelihood phylogenetic analyses were performed on a 594 nt fragment of the ORF1/RdRp region from 84 sinaivirus sequences, including 31 new Lake Sinai viruses (LSVs) from honey bees collected in five countries across the globe and the four ant viral sequences. The phylogeny revealed four main clades potentially representing different viral species infecting honey bees. Moreover, the ant viruses belonged to the LSV4 clade, suggesting a possible cross-species transmission between bees and ants. Lastly, wide honey bee screening showed that all four LSV clades have worldwide distributions with no obvious geographical segregation.