Search

Pollen monitoring have become data-intensive in recent years as real-time detectors are deployed to classify airborne pollen grains. Machine learning models with a focus on deep learning, have an essential role in the pollen classifcation task. Within this study we developed an explainable framework to unveil a deep learning model for pollen classifcation. Model works on data coming from single particle detector (Rapid-E) that records for each particle optical fngerprint with scattered light and laser induced fuorescence. Morphological properties of a particle are sensed with the light scattering process, while chemical properties are encoded with fuorescence spectrum and fuorescence lifetime induced by high-resolution laser. By utilizing these three data modalities, scattering, spectrum, and lifetime, deep learning-based models with millions of parameters are learned to distinguish diferent pollen classes, but a proper understanding of such a black-box model decisions demands additional methods to employ. Our study provides the frst results of applied explainable artifcial intelligence (xAI) methodology on the pollen classifcation model. Extracted knowledge on the important features that attribute to the predicting particular pollen classes is further examined from the perspective of domain knowledge and compared to available reference data on pollen sizes, shape, and laboratory spectrofuorometer measurements.

Varroa destructor is a parasitic mite responsible for the loss of honey bee (Apis mellifera) colonies. This study aimed to find a promising marker in honey for the bee colony infestation level using fluorescence spectroscopy and biochemical analyses. We examined whether the parameters of the honey samples’ fluorescence spectra and biochemical parameters, both related to proteins and phenolics, may be connected with the level of honey bee colonies’ infestation. The infestation level was highly positively correlated with the catalase activity in honey (r = 0.936). Additionally, the infestation level was positively correlated with the phenolic spectral component (r = 0.656), which was tentatively related to the phenolics in honey. No correlation was found between the diastase activity in honey and the colonies’ infestation level. The results indicate that the catalase activity in honey and the PFC1 spectral component may be reliable markers for the V. destructor infestation level of the colonies. The obtained data may be related to the honey yield obtained from the apiaries.

Pollen exposure weakens the immunity against certain seasonal respiratory viruses by diminishing the antiviral interferon response. Here we investigate whether the same applies to the pandemic severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which is sensitive to antiviral interferons, if infection waves coincide with high airborne pollen concentrations. Our original hypothesis was that more airborne pollen would lead to increases in infection rates. To examine this, we performed a cross-sectional and longitudinal data analysis on SARS-CoV-2 infection, airborne pollen, and meteorological factors. Our dataset is the most comprehensive, largest possible worldwide from 130 stations, across 31 countries and five continents. To explicitly investigate the effects of social contact, we additionally considered population density of each study area, as well as lockdown effects, in all possible combinations: without any lockdown, with mixed lockdown−no lockdown regime, and under complete lockdown. We found that airborne pollen, sometimes in synergy with humidity and temperature, explained, on average, 44% of the infection rate variability. Infection rates increased after higher pollen concentrations most frequently during the four previous days. Without lockdown, an increase of pollen abundance by 100 pollen/m3 resulted in a 4% average increase of infection rates. Lockdown halved infection rates under similar pollen concentrations. As there can be no preventive measures against airborne pollen exposure, we suggest wide dissemination of pollen−virus coexposure dire effect information to encourage high-risk individuals to wear particle filter masks during high springtime pollen concentrations.

Background: Information about airborne pollen concentrations is required by a range of end users, particularly from the health sector who use both observations and forecasts to diagnose and treat allergic patients. Manual methods are the standard for such measurements but, despite the range of pollen taxa that can be identified, these techniques suffer from a range of drawbacks. This includes being available at low temporal resolution (usually daily averages) and with a delay (usually 3–9 days from the measurement). Recent technological developments have made possible automatic pollen measurements, which are available at high temporal resolution and in real time, although currently only scattered in a few locations across Europe. Materials & Methods: To promote the development of an extensive network across Europe and to ensure that this network will respond to end user needs, a stakeholder workshop was organised under the auspices of the EUMETNET AutoPollen Programme. Participants discussed requirements for the groups they represented, ranging from the need for information at various spatial scales, at high temporal resolution, and for targeted services to be developed. Results: The provision of real-time information is likely to lead to a notable decrease in the direct and indirect health costs associated with allergy in Europe, currently estimated between €50–150 billion/year.. Discussion & Conclusion: A European measurement network to meet end user requirements would thus more than pay for itself in terms of potential annual savings and provide significant impetus to research across a range of disciplines from climate science and public health to agriculture and environmental management.

Background: Information about airborne pollen concentrations is required by a range of end users, particularly from the health sector who use both observations and forecasts to diagnose and treat allergic patients. Manual methods are the standard for such measurements but, despite the range of pollen taxa that can be identified, these techniques suffer from a range of drawbacks. This includes being available at low temporal resolution (usually daily averages) and with a delay (usually 3–9 days from the measurement). Recent technological developments have made possible automatic pollen measurements, which are available at high temporal resolution and in real time, although currently only scattered in a few locations across Europe. Materials & Methods: To promote the development of an extensive network across Europe and to ensure that this network will respond to end user needs, a stakeholder workshop was organised under the auspices of the EUMETNET AutoPollen Programme. Participants discussed requirements for the groups they represented, ranging from the need for information at various spatial scales, at high temporal resolution, and for targeted services to be developed. Results: The provision of real-time information is likely to lead to a notable decrease in the direct and indirect health costs associated with allergy in Europe, currently estimated between €50–150 billion/year.1 Discussion & Conclusion: A European measurement network to meet end user requirements would thus more than pay for itself in terms of potential annual savings and provide significant impetus to research across a range of disciplines from climate science and public health to agriculture and environmental management., Peer Reviewed, Postprint (published version)

Common ragweed (Ambrosia artemisiifolia L.) is a widely distributed and harmful invasive plant that is an important source of highly allergenic pollen grains and a prominent crop weed. As a result, ragweed causes huge costs to both human health and agriculture in affected areas. Efficient mitigation requires accurate mapping of ragweed densities that, until now, has not been achieved accurately for the whole of Europe. Here we provide two inventories of common ragweed abundances with grid resolutions of 1 km and 10 km. These “top-down” inventories integrate pollen data from 349 stations in Europe with habitat and landscape management information, derived from land cover data and expert knowledge. This allows us to cover areas where surface observations are missing. Model results were validated using “bottom– up” data of common ragweed in Austria and Serbia. Results show high agreement between the two analytical methods. The inventory shows that areas with the lowest ragweed abundances are found in Northern and Southern European countries and the highest abundances are in parts of Russia, parts of Ukraine and the Pannonian Plain. Smaller hotspots are found in Northern Italy, the Rhône Valley in France and in Turkey. The top-down approach is based on a new approach that allows for cross- continental studies and is applicable to other anemophilous species. Due to its simplicity, it can be used to investigate such species that are difficult and costly to identify at larger scales using traditional vegetation surveys or remote sensing. The final inventory is open source and available as a georeferenced tif file, allowing for multiple usages, reducing costs for health services and agriculture through well- targeted management interventions.

The recent decline in managed honey bee populations, Apis mellifera L. (Hymenoptera: Apidae), has caused scientific, ecological, and economic concern. Research into the formation of reactive oxygen species (ROS), antioxidative defense mechanisms, and oxidative stress can contribute to our understanding of bee survival and conservation of this species. Activities of superoxide dismutase (SOD), catalase (CAT), and glutathione S-transferase (GST) enzymes together with levels of malondialdehyde (MDA) were measured in summer and winter honey bees sampled from three colonies. One colony was stationary (C1), entering the winter period having accumulated Robinia pseudoacacia L. (Fabaceae) honey, and two were migratory (C2 and C3), entering the winter period with mainly Tilia (Malvaceae) and Brassica (Brassicaceae) honey, respectively. Compared to summer workers, winter worker bees had decreased SOD and GST activity, and MDA level, whereas CAT activity increased in all three colonies. We also demonstrated that seasonality is the main factor responsible for changes in antioxidant enzymes and MDA levels in worker honey bees. Overall, our results indicate a difference between summer and winter worker bees, pointing at a reduced level of antioxidant enzyme defenses during overwintering which may be due to a decrease in production of ROS. The decreased levels of MDA measured in winter honey bees confirm this. As ROS are actively used by insects as a defense mechanism to fight pathogens, we suggest that reduced production of ROS contributes to higher susceptibility of winter honey bees to infections and reduced overwinter survival.

Fluorescence is non-destructive, sensitive, simple and fast method for analysis of fluorescent compounds contained in very low amounts (nanomolar concentrations) in the samples. It can be used for structural or concentration studies, in analytical or diagnostic purposes [1].The fluorescence spectra, in combination with appropriate statistical methods, may provide useful fingerprints in food analysis [2]. Various methods for study of honey quality and adulteration have been in research focus [3]. Over the last years, in different geographic areas a notable los of honey bee (Apis mellifera L.) colonies has been reported. A number of stressors affecting honey bees, including diseases, parasites, pesticides and poor nutrition have been identified [4]. Therefore fast and reliable methods are required for screening bee products both as a tool for assessing quality and to identify risks for colony state. We used fluorescence spectroscopy combined with advanced statistical analysis in order to identify variability in Fruska Gora lime tree (Tilia L.) honey collected at different locations in 2015. Since homogenization of the honey before packing in jars is considered as critical procedure from the Quality Control point of view, we have explored to what extent the ratio of the two main fluorophores in honey, originating from proteins and phenolic compounds change between extraction stage to packaging. Steady state fluorescence spectroscopy in combination with Multivariate Curve Resolution Alternating Least Squares (MCR-ALS) for spectral analysis has been applied to differentiate samples of honey. The three-dimensional excitation–emission matrix (EEM) is a rapid, selective and sensitive method: by changing excitation and the emission wavelength simultaneously, information regarding the fluorescence characteristics of the different compounds contained in the sample of interest can be obtained [5]. Proteins in honey mainly originate from bees and their quantity depends on bee so

Physicochemical composition and techno-functional properties of bee pollens collected in Serbia were assessed. Analysed bee pollen contained 14.81-27.25% proteins, 1.31-6.78% lipids, 64.42-81.84% carbohydrates and 1.18-3.21% ash, with mean energy value of 375 kcal. Bee pollen showed low protein solubility (2.79-25.90 g/100 g), high carbohydrate solubility (31.2-75 g/100 g), good emulsifying properties (emulsion stability index ranged from 19.6 to 49.3 min and emulsion activity index ranged from 10.40 to 24.52 m(2)/g), non-foaming properties, poor water absorption capacity (0.92-2.25 g/g) and excellent oil absorption capacity (1-3.53 g/g). Protein solubility was positively correlated with carbohydrate content (r = 0.73, p lt 0.05), but negatively with ash and lipid content (r = -0.39, r = -0.46, p lt 0.05, respectively). The total protein content and lipid content were shown positive relationship with carbohydrate solubility (r = 039, r = 0.45, p lt 0.05, respectively). Emulsion stability was positively correlated with protein solubility (r = 0.47, p lt 0.05), whereas emulsion activity was negatively correlated with this parameter (r = -0.39, p lt 0.05). Water and oil absorption capacity were not shown significant correlations with other investigated parameters. The obtained data indicated that bee pollen could find useful application as food ingredient in variety of food products.

Understanding the cellular stress response in honey bees will significantly contribute to their conservation. The aim of this study was to analyze the response of the antioxidative enzymes superoxide dismutase and catalase in honey bees related to the presence of toxic metals in different habitats. Three locations were selected: (i) Tunovo on the mountain Golija, as control area, without industry and large human impact, (ii) Belgrade as urban area, and (iii) Zajaca, as mining and industrial zone. Our results showed that the concentrations of lead (Pb) in whole body of bees vary according to habitat, but there was very significant increase of Pb in bees from investigated industrial area. Bees from urban and industrial area had increased expression of both Sod1 and Cat genes, suggesting adaptation to increased oxidative stress. However, in spite increased gene expression, the enzyme activity of catalase was lower in bees from industrial area suggesting inhibitory effect of Pb on catalase.