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.

Epidemiology of a major honey bee pathogen, deformed wing virus: potential worldwide replacement of genotype A by genotype B.

The western honey bee (Apis mellifera) is of major economic and ecological importance, with elevated rates of colony losses in temperate regions over the last two decades thought to be largely caused by the exotic ectoparasitic mite Varroa destructor and deformed wing virus (DWV), which the mite transmits. DWV currently exists as two main genotypes: the formerly widespread DWV-A and the more recently described and rapidly expanding DWV-B. It is an excellent system to understand viral evolution and the replacement of one viral variant by another. Here we synthesise published results on the distribution and prevalence of DWV-A and -B over the period 2008-2021 and present novel data for Germany, Italy and the UK to suggest that (i) DWV-B has rapidly expanded worldwide since its first description in 2004 and (ii) that it is potentially replacing DWV-A. Both genotypes are also found in wild bee species. Based on a simple mathematical model, we suggest that interference between viral genotypes when co-infecting the same host is key to understanding their epidemiology. We finally discuss the consequences of genotype replacement for beekeeping and for wild pollinator species.

Molecular Detection of Malpighamoeba mellificae in Honey Bees.

Malpighamoeba mellificae is a protozoan that infects the Malpighian tubules of honey bees. The amoebae, ingested as cysts, develop into trophozoites that feed upon tubule epithelia. The resulting damage of the Malpighian tubules can induce an imbalance of waste excretion and hemolymph exchange. This causes the so-called amoebiasis disease in adult bees, which may co-occur with Nosema infections. Most reports of this amoeba are from the 1960s and earlier, and knowledge of the disease and its spreading is very poor. The lack of any genetic marker for the species hampers its sensitive identification using molecular tools and gaining knowledge on its epidemiology. Here, we present a diagnostic RT-qPCR assay, consisting of two primers and one probe that were developed based on 18S rRNA sequences of the amoeba, generated with metagenomic sequencing of Malpighian tubules with and without M. mellificae cysts. The assay was initially tested and adjusted with samples microscopically tested for the presence of M. mellificae cysts. Later, it was validated and material with unknown infection status was tested. The sensitive diagnostic Malpighamoeba disease 18S assay is now ready to be applied for honey bee health monitoring purposes and to investigate the prevalence of M. mellificae in more detail.

Development and evaluation of a core genome multilocus sequence typing scheme for Paenibacillus larvae, the deadly American foulbrood pathogen of honeybees.

Paenibacillus larvae is the causative agent of the fatal American foulbrood disease in honeybees (Apis mellifera). Strain identification is vital for preventing the spread of the disease. To date, the most accessible and robust scheme to identify strains is the multilocus sequence typing (MLST) method. However, this approach has limited resolution, especially for epidemiological studies. As the cost of whole-genome sequencing has decreased and as it becomes increasingly available to most laboratories, an extended MLST based on the core genome (cgMLST) presents a valuable tool for high-resolution investigations. In this study, we present a standardized, robust cgMLST scheme for P. larvae typing using whole-genome sequencing. A total of 333 genomes were used to identify, validate and evaluate 2419 core genes. The cgMLST allowed fine-scale differentiation between samples that had the same profile using traditional MLST and allowed for the characterization of strains impossible by MLST. The scheme was successfully used to trace a localized Swedish outbreak, where a cluster of 38 isolates was linked to a country-wide beekeeping operation. cgMLST greatly enhances the power of a traditional typing scheme, while preserving the same stability and standardization for sharing results and methods across different laboratories.

Veterinary Diagnostic Approach of Common Virus Diseases in Adult Honeybees.

Veterinarians are educated in prevention, diagnosis and treatment of diseases in various vertebrate species. As they are familiar with multifactorial health problems in single animals as well as in herd health management, their knowledge and skills can be beneficial for the beekeepers and honeybee health. However, in education and in practice, honeybees are not a common species for most veterinarians and the typical veterinary diagnostic methods such as blood sampling or auscultation are not applicable to the superorganism honeybee. Honeybee colonies may be affected by various biotic and abiotic factors. Among the infectious agents, RNA-viruses build the largest group, causing covert and overt infections in honeybee colonies which may lead to colony losses. Veterinarians could and should play a more substantial role in managing honeybee health-not limited to cases of notifiable diseases and official hygiene controls. This review discusses the veterinary diagnostic approach to adult bee examination with a special focus on diagnosis and differential diagnosis of the common virus diseases Acute Bee Paralysis Virus (ABPV)-Kashmir Bee Virus (KBV)-Israeli Acute Paralysis Virus (IAPV)-Complex, Chronic Bee Paralysis Virus (CBPV) and Deformed Wing Virus (DWV), as well as coinfections like Varroa spp. and Nosema spp.

Overt Infection with Chronic Bee Paralysis Virus (CBPV) in Two Honey Bee Colonies.

Chronic Bee Paralysis Virus (CBPV), a widespread honey bee RNA virus, causes massive worker bee losses, mostly in strong colonies. Two different syndromes, with paralysis, ataxia and flight incapacity on one hand and black hairless individuals with shortened abdomens on the other, can affect a colony simultaneously. This case report presents two Apis mellifera carnica colonies with symptoms of paralysis and hairless black syndrome in 2019. Via RT-PCR, a highly positive result for CBPV was detected in both samples. Further problems, such as a Nosema infection and Varroa infestation, were present in these colonies. Therapy methods were applied to colony 1 comprising queen replacement, shook swarm method and Varroa control, whereas colony 2 was asphyxiated after queen loss and colony weakening. After therapy, colony 1 was wintered without symptoms. Beekeeping and sanitary measures can save a CBPV-infected colony, while further complications result in total colony loss.

Trueness and precision of the real-time RT-PCR method for quantifying the chronic bee paralysis virus genome in bee homogenates evaluated by a comparative inter-laboratory study.

The Chronic bee paralysis virus (CBPV) is the aetiological agent of chronic bee paralysis, a contagious disease associated with nervous disorders in adult honeybees leading to massive mortalities in front of the hives. Some of the clinical signs frequently reported, such as trembling, may be confused with intoxication syndromes. Therefore, laboratory diagnosis using real-time PCR to quantify CBPV loads is used to confirm disease. Clinical signs of chronic paralysis are usually associated with viral loads higher than 108 copies of CBPV genome copies per bee (8 log10 CBPV/bee). This threshold is used by the European Union Reference Laboratory for Bee Health to diagnose the disease. In 2015, the accuracy of measurements of three CBPV loads (5, 8 and 9 log10 CBPV/bee) was assessed through an inter-laboratory study. Twenty-one participants, including 16 European National Reference Laboratories, received 13 homogenates of CBPV-infected bees adjusted to the three loads. Participants were requested to use the method usually employed for routine diagnosis. The quantitative results (n=270) were analysed according to international standards NF ISO 13528 (2015) and NF ISO 5725-2 (1994). The standard deviations of measurement reproducibility (SR) were 0.83, 1.06 and 1.16 at viral loads 5, 8 and 9 log10 CBPV/bee, respectively. The inter-laboratory confidence of viral quantification (+/- 1.96SR) at the diagnostic threshold (8 log10 CBPV/bee) was+/- 2.08 log10 CBPV/bee. These results highlight the need to take into account the confidence of measurements in epidemiological studies using results from different laboratories. Considering this confidence, viral loads over 6 log10 CBPV/bee may be considered to indicate probable cases of chronic paralysis.

Imaging live bee brains using minimally-invasive diagnostic radioentomology.

The sensitivity of the honey bee, Apis mellifera L. (Hymeonoptera: Apidae), brain volume and density to behavior (plasticity) makes it a great model for exploring the interactions between experience, behavior, and brain structure. Plasticity in the adult bee brain has been demonstrated in previous experiments. This experiment was conducted to identify the potentials and limitations of MicroCT (micro computed tomograpy) scanning "live" bees as a more comprehensive, non-invasive method for brain morphology and physiology. Bench-top and synchrotron MicroCT were used to scan live bees. For improved tissue differentiation, bees were fed and injected with radiographic contrast. Images of optic lobes, ocelli, antennal lobes, and mushroom bodies were visualized in 2D and 3D rendering modes. Scanning of live bees (for the first time) enabled minimally-invasive imaging of physiological processes such as passage of contrast from gut to haemolymph, and preliminary brain perfusion studies. The use of microCT scanning for studying insects (collectively termed 'diagnostic radioentomology', or DR) is increasing. Our results indicate that it is feasible to observe plasticity of the honey bee brain in vivo using diagnostic radioentomology, and that progressive, real-time observations of these changes can be followed in individual live bees. Limitations of live bee scanning, such as movement errors and poor tissue differentiation, were identified; however, there is great potential for in-vivo, non-invasive diagnostic radioentomology imaging of the honey bee for brain morphology and physiology.

Winter losses of honeybee colonies (Hymenoptera: Apidae): the role of infestations with Aethina tumida (Coleoptera: Nitidulidae) and Varroa destructor (Parasitiformes: Varroidae).

Multiple infections of managed honeybee, Apis mellifera, colonies are inevitable due to the ubiquitous ectoparasitic mite Varroa destructor and might be an underlying cause of winter losses. Here we investigated the role of adult small hive beetles, Aethina tumida, alone and in combination with V. destructor for winter losses and for infections with the microsporidian endoparasite Nosema ceranae. We found no significant influence of A. tumida and V. destructor alone or in combination on the numbers of N. ceranae spores. Likewise, A. tumida alone had no significant effects on winter losses, which is most likely due to the observed high winter mortality of the adult beetles. Therefore, our data suggest that A. tumida is unlikely to contribute to losses of overwintering honeybee colonies. However, high losses occurred in all groups highly infested with V. destructor, supporting the central role of the mite for colony losses.