Progress 01/01/24 to 12/31/24
Outputs
Target Audience:Dr. Anderson: Calgary and District Beekeepers Association. (Invited presentation) November 27, 2024 at 7pm Kirk Anderson, a bee microbial ecologist, will present on Probiotics/Antibiotics. Anderson K.E. April 5, 2024. (Invited presentation) Apiaries and Bees for Communities; Master Academy webinar series; "An Inside Job: The inner-workings of the honey bee colony: Honey Bee Microbial Ecology" 2024 Nov. Invited interview. Microbiome research presented on "Inside the Hive TV-Podcast" organized by Humberto Boncristiani "Are Probiotics Really Helping Bees?" https://www.youtube.com/playlist?app=desktop&list=PLzMvo1IUOsOfEUEnPVOm83KlfKWSgi8QM 2024 Sept. Invited interview. Microbiome research presented on "Two Bees in a Podcast, organized by University of Florida Honey bee lab Jamie Ellis and Amy Wu "Commercial Honey bees and Non-native Probiotics"https://open.spotify.com/episode/5F5Z4uWoJGk623tVy7auAr 2024 July. Invited Article. Bee Culture, "The hopeful state of probiotic science"https://www.beeculture.com/the-hopeful-state-of-probiotic-science/ 2024 May. Microbiome research featured in American Bee Journal; "A Field Trial of Probiotics: Part 2" by Randy Oliverhttps://scientificbeekeeping.com/8601-2/ Dr. Copeland: Aug 2024 Honey Bee Microbial Ecology in the Age of Bioinformatics and Data Science (Davis, CA) Sep 2024 USDA Pollinator Lightning Talk (remote) Nov 2024 USDA ARS AI Forum Lightning Talk (College Station, TX) Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?One postdoc scientist, two graduate students, ≥3technicians, and ≥3 undergraduate students at CHBRC, MSU, and BRL have been trained on various micro- and molecular-biology techniques, general apiology, and bioinformatic techniques for this project. These include: Bacterial isolation, molecular and metabolic characterization, PCR screening and sequencing for strain typing and virulence marker discovery, and in-vitro studies of inhibition and virulence in honey bee larvae. Graduate students and professional staff have also been developing various informatic skills for novel (virulence) marker discovery, microbiome and genomic data analysis, and machine-learning (AI) development. Our efforts to build a broader national network of collaborators have provided mutual professional development for our postdoc, project manager, and various apiary inspectors, beekeepers and extension agents as we seek to educate and encourage their participation in this project. How have the results been disseminated to communities of interest?In 2024 we published several papers relating to Obj.1: 1) A large field study on use of antibiotics and probiotics in a commercial operation, which supports a post-antibiotic dysbiosis observation in our IL pilot study published 2023 and generally confirms the risks of antibiotic use as may be related to misdiagnosis of EFB and other infections. (https://doi.org/10.1038/s41598-024-52118-z). 2) A queen microbiome study that complements the nexus of queen-larvae-worker microbiomes, pathogen infection, and individual immunity as above (https://doi.org/10.1038/s41598-024-58883-1). 3) The largest-ever meta-analysis of honey bee microbiome studies, with insights on the role of a "social resource microbiota" in disease prevention in worker larvae, adult workers, and adult queens. (https://doi.org/10.3389/frbee.2024.1410331 ). We have several other in-prep publications relating to Obj.1 that will be published in 2025: 1) The largest-ever survey of larval microbiomes in health and diseased (noted above). 2) The prevalence, abundance and co-occurrence of bacterial (M.plutonius) and viral pathogens in honey bee hives presenting with symptoms of EFB-like disease. 3) An even larger meta-analysis of honey bee gut microbiota with an emphasis on defining the rare biosphere bacterial species that may contribute to gut resilience and susceptibility. We've also presented our findings at numerous professional meetings, including: Dr. Anderson: Calgary and District Beekeepers Association. (Invited presentation) November 27, 2024 at 7pm Kirk Anderson, a bee microbial ecologist, will present on Probiotics/Antibiotics. Anderson K.E. April 5, 2024. (Invited presentation) Apiaries and Bees for Communities; Master Academy webinar series; "An Inside Job: The inner-workings of the honey bee colony: Honey Bee Microbial Ecology" 2024 Nov. Invited interview. Microbiome research presented on "Inside the Hive TV-Podcast" organized by Humberto Boncristiani "Are Probiotics Really Helping Bees?" https://www.youtube.com/playlist?app=desktop&list=PLzMvo1IUOsOfEUEnPVOm83KlfKWSgi8QM 2024 Sept. Invited interview. Microbiome research presented on "Two Bees in a Podcast, organized by University of Florida Honey bee lab Jamie Ellis and Amy Wu "Commercial Honey bees and Non-native Probiotics"https://open.spotify.com/episode/5F5Z4uWoJGk623tVy7auAr 2024 July. Invited Article. Bee Culture, "The hopeful state of probiotic science"https://www.beeculture.com/the-hopeful-state-of-probiotic-science/ 2024 May. Microbiome research featured in American Bee Journal; "A Field Trial of Probiotics: Part 2" by Randy Oliverhttps://scientificbeekeeping.com/8601-2/ Dr. Copeland: Aug 2024 Honey Bee Microbial Ecology in the Age of Bioinformatics and Data Science (Davis, CA) Sep 2024 USDA Pollinator Lightning Talk (remote) Nov 2024 USDA ARS AI Forum Lightning Talk (College Station, TX) What do you plan to do during the next reporting period to accomplish the goals?Objectives 1 and 4) Determine the microbiome associated with larval health and disease, and Improve the detection and diagnosis of larval disease, both known and discovered. Through our collaboration with the Apiary Inspectors of America, we continue to expand collections of a wide variety of brood disease images and corresponding disease microbiomes. To produce our Ai diagnostic tool, we are collecting the needed depth of information (images) from non-EFB disease states, including Varrosis or Parasitic Mite Syndrome (PMS) and Idiopathic Brood Disease Syndrome (IBDS). To provide the variation needed to increase the confidence of our predictive tool, we will expand our AI-training set to include greater representation of non-EFB larval disease symptomology from throughout the US. Based on past results, we calculate that ten more sampling events of this size and character are needed to complete the training set for our digital tool. We have recruited multiple state apiary inspectors to take digital images of brood disease phenotypes and ship diseased brood samples directly to our lab for molecular diagnostic screening and subsequent microbiome analysis. We are applying machine learning to predict diseased microbiomes based on digital imagery, and our collaboration with the apiary inspectors provides the detailed images to further reinforce our training set. The postdoctoral researcher continues to develop and test a Convolutional Neural Network (CNN) with SCINet resources to extract critical features from high-resolution digital images of larval disease phenotypes that possess corresponding microbiome data. Because different brood diseases are associated with distinct symptomology, our present CNN can identify disease agents with a high degree of accuracy following training with a set of informed examples. Ultimately, this work will provide a new tool for beekeepers and apiary inspectors to quickly access image based diagnostic services. Objective 2) We will continue the in vitro testing focused on a variety of non-EFB opportunists identified in the microbiome screen. As an added dimension, we are collaborating with Beltsville diagnostic service Dr. Mohamed Alburaki) to produce a beekeeper specific application to diagnose disease in the field and provide other important metrics associated with the pollination landscape. Objective 3) We continue to determine the distribution of Melissococcus clonal groups and virulence genes within apiaries across landscapes, and throughout the US. We have amassed the DNA primer sets and geographic samples sets to begin this monumental task, and lab work has begun, guided by a molecular screening for virulence.
Impacts
What was accomplished under these goals?
General: In anticipation of the grant's end, we've slowed sample collection to focus on project deliverables in two forms: 1) sample processing and data analysis for research publications, and 2) exploring new partnerships to ensure the long-term accessibility and utility of our work for both research and ag-industry applications. Nevertheless, we expanded our collaborator network and more than doubled our brood disease photo database linked to molecularly-diagnosed hives in 2024. We now have support from individuals or groups in 16 states assisting with brood disease sample collection (i.e., State Ag. Dept. apiary inspectors, University Extension agents, and individual beekeepers in MA, WI, OR, WV, NJ, NH, OH, HI, CA, AZ, CO, IL, MI, PA, NY, MD).). Combined with USDA-ARS-BRL submissions, we have EFB, EFB-like, Parasitic mite syndrome (PMS), and idiopathic brood disease (IBDS) samples from 28 U.S. states. We've also continued to improve our hive-level disease screening and bulk photo collection protocol for AI training. This yields more AI-ready collections from more hives and operations over more states. Our molecular diagnostic and bacterial isolate screening workflows are ongoing. We've run qPCR on >1000 larvae from >50 hives, screening for 10 viruses, EFB, and total bacteria. Hundreds of bacterial isolates have also been screened to identify M. plutonius strains and other taxa associated with larval health and disease. Objective 1: Our first publication for Obj.1 (pilot study on 6 IL apiaries) revealed microbiome associations, developmental succession, and opportunistic pathogens in five hives afflicted with EFB (M. plutonius). Notably, the sixth hive was diagnosed as--and presented with many symptoms of--EFB, but molecular analysis revealed a disease state linked to Acute Bee Paralysis Virus and various opportunistic bacteria (https://doi.org/10.1038/s41598-023-28085-2). We've been steadily building on this design with similar collections across multiple states. In 2024 we completed 16S rDNA microbiome sequencing on N=616 larvae from 8 apiaries in MI. These were added to ~20 new IL and 2 AZ/CA sites to create the largest ever survey of larval gut microbiota diversity (>1400 individuals from healthy and diseased hives). The initial survey should be published in 2025, but these microbiomes are linked to other data (photos, RNA, live bacterial isolates, etc.), ensuring we maximize the utility of each sample. We now have 4 years of photo + larval collection from co-PD Milbrath, plus smaller collections from six other states in various stages of analysis. . In 2024, >1000 larvae were screened for EFB and 10 viruses via qPCR, ~576 are queued for 16S rDNA sequencing, and 282 larvae from 3 sites were analyzed via qPCR for a study of larval immune gene expression. Objectives 2 and 3: We collected >100 new bacterial isolates in 2024. Over the life of the project, we've cultured nearly 1400 bacterial isolates and confirmed >400 as M. plutonius. Among these, we've identified ≥17 unique sequence types and ≥7 new alleles in two clonal complexes (CC3 and CC12) utilizing published MLST markers. Combined with variation in mtxa positivity (a putative virulence gene on pMP19-plasmid) and several other markers, we likely have several dozen unique strains ready for genomic sequencing and further study. To characterize their virulence, >30 of these M.plutonius isolates were tested in dose-response studies on in-vitro reared worker larvae. Most strains were highly virulent (i.e., more lethal than strains studied with equivalent methods in our previous publications), but a few were apparently benign (i.e., survival and development rates equal to control larvae inoculated with normal saline). We also confirmed bacterial inoculum concentration and survival in royal jelly and observed variable growth rates on agar media, broth, etc. In 2024 we implemented various changes to these in-vitro larval studies to increase dosing precision and biological insights, and we've identified further differences in patterns of M.plutonius infection and in-vivo growth that we plan to link with comparative genomic studies in 2025. These projects will complement our previously described PAMPs study (pathogen-associated molecular patterns), wherein N=700 24-hour-old larvae were assigned one of 8 treatments (2 controls plus 6 diets composed of either: gram-positive bacteria (M. plutonius), gram-negative bacteria (Serratia sp.), or one of four PAMPs derived from gram-positive bacteria, gram-negative bacteria, and fungi). These PAMPs are known to trigger immune and oxidative stress responses in other animal systems, so we sampled larvae every 24 hours and measured their gene expression via qPCR for 13 markers associated with innate immunity and stress in honey bees. Together, these projects provide a thorough examination of bacterial pathogen diversity, virulence mechanisms, and larval immune response. We anticipate at least preliminary results from this work will be published in 2025. Objective 4: All of the work detailed above contributes directly and indirectly to our knowledge, detection, diagnosis and even potential treatments for honey bee disease. Specific contributions to diagnosis include both our optimized, pooled-sample workflows for molecular diagnostics and the complementary AI-image-diagnosis tool. Our pooled-sample qPCR approach has been adapted from methods introduced for mass viral-pathogen qPCR testing. This has allowed us to screen hundreds of individual samples in pools (each composed of biologically similar replicates, e.g., 3rd instar healthy larvae), and then target our individual-sample qPCR on the groups and markers where pooled signals were significant. These molecular diagnoses are paired with a database that now includes >8,000 photos of diseased and healthy brood. These molecularly-diagnosed photos create an extremely high-quality training set, which is the basis for our rapid advances in creating an AI-based application to diagnose honey bee brood diseases via smart-phone photography. This BDBD-AI application will allow us to quickly and accurately identify different diseases based on visual characteristics of the larvae. By using machine learning techniques, we have built a working model on a small subset of the data and confirmed its ability to distinguish EFB from PMS-associated viral symptomologies. As detailed above, misdiagnosis of EFB and EFB-like PMS is a widespread problem observed in our data and results in significant costs to honey bees, beekeepers, and agriculture as a whole. The application will also be able to make recommendations for treatment based on the identified disease. This will be a valuable tool for beekeepers, researchers, and other professionals working with bees, as it will enable them to quickly and accurately diagnose and treat diseases in their hives. In 2024 we focused on increasing the training set for this tool and more than doubled our number of photos paired with molecular diagnosis. We also met with various external and internal (USDA) stakeholders, web and mobile-phone app developers, and initiated new partnerships with several USDA collaborators that can provide long-term storage of our burgeoning database. We will expand this work in 2025, including collaboration with Dr. Alburaki of the USDA-BRL on efforts to modernize their bee diagnostic service and working with members of the USDA Pollinator Data Portal to ensure our work remains accessible and extensible for future growth.
Publications
- Type:
Peer Reviewed Journal Articles
Status:
Accepted
Year Published:
2024
Citation:
Anderson KE and Copeland DC (2024) The honey bee hive microbiota: meta-analysis reveals a native and aerobic microbiota prevalent throughout the social resource niche. Front. Bee Sci. 2:1410331. doi: 10.3389/frbee.2024.1410331
|
Progress 01/01/23 to 12/31/23
Outputs
Target Audience:Target audiences reached in 2022 include the American Honey Producers Association - Annual Conference and Trade Show in Tucson, AZ, in December 2022. We have been invited to present our findings to the Apiary Inspectors of America held in conjunction with the American Beekeeping Federation concerning recent progress and sampling details needed to further pursue Image-Based-Diagnostic-Services for honey bee brood disease. We have solidified an agreement with corperate sponsor Mann Lake Ltd. to produce asmartphone application that can diagnose the cause of brood disease in the field, either viral or bacterial, reducing the unnecessary use of antibiotics. Mann Lake Ltd. has invited our team to speak at the upcoming national meetings in 2024. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?One postdoc scientist, four graduate students, ≥4 technicians, and ≥4 undergraduate students at CHBRC, MSU, and BRL have been trained on various micro- and molecular-biology, apiology, and bioinformatic techniques for this project. These include: Bacterial isolation, molecular and metabolic characterization, PCR screening and sequencing for strain typing and virulence marker discovery, and in-vitro studies of inhibition and virulence in honey bee larvae. Graduate students and professional staff have also been developing various informatic skills for novel (virulence) marker discovery, microbiome and genomic data analysis, and machine-learning (AI) development. Our efforts to build a broader national network of collaborators have provided mutual professional development for our postdoc, project manager, and various apiary inspectors, beekeepers and extension agents as we seek to educate and encourage their participation in this project. How have the results been disseminated to communities of interest?In January we published the largest larval microbiome dataset to date (Jan. 2023, https://doi.org/10.1038/s41598-023-28085-2 , detailed in accomplishments above). This paper, other recent findings on brood disease microbiomes, misdiagnosis of cryptic viral infections, and a call for sample submissions to further our work on diagnostics via AI-image recognition, were all presented to the Apiary Inspectors of America in (May 10th), highlighted in Catch The Buzz (May) and Bee Culture Magazine (Aug.), and disseminated via numerous smaller meetings with State Apiarists, University Extension Agents, and private beekeepers in six states (May-Aug). We've also held meetings with several parties interested in establishing a USDA-ARS-CRADA for the commercial development and dissemination of our AI-image-diagnosis tool. What do you plan to do during the next reporting period to accomplish the goals?Objectives 1 and 4) Determine the microbiome associated with larval health and disease, and Improve the detection and diagnosis of larval disease, both known and discovered. Through our collaboration with the Apiary Inspectors of America, we are collecting a wide variety of brood disease images and corresponding disease microbiomes. We lack the needed depth of information (images) from non-EFB disease states, including Varrosis or Parasitic Mite Syndrome (PMS) and Idiopathic Brood Disease Syndrome (IBDS). To provide the variation needed to increase the confidence of our predictive tool, we will expand our AI-training set to include greater representation of non-EFB larval disease symptomology from throughout the US. Based on past results, we calculate that ten more sampling events of this size and character are needed to complete the training set for our digital tool. We will recruit approximately ten state apiary inspectors to take digital images of brood disease phenotypes and ship diseased brood samples directly to our lab for molecular diagnostic screening and subsequent microbiome analysis. We are applying machine learning to predict diseased microbiomes based on digital imagery, and our collaboration with the apiary inspectors provides the detailed images to further reinforce our training set. The postdoctoral researcher is presently developing a Convolutional Neural Network (CNN) with SCINet resources to extract critical features from high-resolution digital images of larval disease phenotypes that possess corresponding microbiome data. CNNs are flexible and have already been successfully implemented to solve many image-based challenges in honey bees and other organisms. Because different brood diseases are associated with distinct symptomology, our present CNN can identify disease agents with a high degree of accuracy following training with a set of informed examples. Ultimately, this work will provide a new tool for beekeepers and apiary inspectors to quickly access image based diagnostic services. Objective 2) We will continue the in vitro testing focused on a variety of non-EFB opportunists identified in the microbiome screen. As an added dimension, we are collaborating with Dalan Health Inc. to understand the inheritance of EFB molecular patterns from queen to larvae and its effect on larval immunity. Objective 3) We continue to determine the distribution of Melissococcus clonal groups and virulence genes within apiaries across landscapes, and throughout the US. We have amassed the DNA primer sets and geographic samples sets to begin this monumental task, and lab work has begun, guided by a molecular screening for virulence.
Impacts
What was accomplished under these goals?
General: To increase sample collection, weinvited stakeholder support through various channels.This led to six new active collaborations with State Ag. Dept. apiary inspectors, University Extension agents, and individual beekeepers in WI, OR, WV, NJ, NH and OH, and at least three more prospective collaborations. Combined with collaborations initiated in prior years (HI, CA, AZ, CO, IL, MI, PA, NY, MD), we now have individuals or groups in 15 states directly assisting us with brood disease sample collection. These are augmented by indirect sample submissions through the USDA-ARS-BRL disease diagnosis service. Together, these networks have provided us with various disease symptomology from 28 U.S. states and growing. We havefully implemented our molecular diagnostic and bacterial isolate-screening workflowsfor 2023 that focused on hive-level screening for disease and bulk photo collection to train the AI for brood disease image recognition. These changes meant fewer samples per hive, but more AI-ready (photos+larvae) collections from more hives and operations over more states. To date we've run >1600 larvae through our qPCR pathogen panel (10 viruses, EFB, and total bacteria) and begun screening hundreds of bacterial isolates(samplesdistributed over 10 states, 22 beekeeping operations, and 37 hives) to identify unique strains of M.plutonius and other taxa associated with larval health and disease. Objective 1: Our first publication for Obj.1, yielded many insights on the complex microbiome associations, developmental succession, and opportunistic pathogens in hives and individual larvae afflicted primarily with M.plutonius. Critically, this study also revealed a case of misdiagnosis where Varroosis can present in a somewhat EFB-like manner, a disease state linked to Acute Bee Paralysis Virus and various opportunistic bacteria (https://doi.org/10.1038/s41598-023-28085-2). This pilot study was the foundation for much of this NIFA grant, is now matched by even larger collections from co-PD Milbrath in MI, and represented just ~1/5 of the nominal EFB apiaries collected in IL. We have now completed 16S rDNA microbiome sequencing and initial bioinformatics on an additional ~20 apiaries from IL, all are matched with high-quality photographs to facilitate AI training. We now have 4 years of intensive photo and larval collection from co-PD Milbrath in various stages of lab processing and analysis: In 2023 we screened ~1500 of these larvae from 2021 and 2022 for EFB and 10 viruses via qPCR, ~576 of them are in queue for 16S rDNA microbiome sequencing via our new (Oct. 2023) high-throughput sequencing contract, and 282 larvae from 3 sites have been further analyzed via qPCR for a study of larval immune gene expression associated with viral and bacterial pathogen titers, healthy and pathogen-perturbed development, and succession of visible disease phenotypes. Preliminary results already indicate several additional apiaries where EFB-like, Idiopathic Brood disease (including melty larvae) is linked to high viral titers and little or no M.plutonius, which substantiates major hypotheses for this project and-we expect-will lead to multiple additional scientific publications on disentangling these disease syndromes. Objectives 2 and 3: In 2023 we collected another 253 bacterial isolates, including 191 putative M.plutonius, from a broad range of sources (detailed above). In total we've collected over 1100 unique isolates from diseased brood, ~770 of which have been identified as M.plutonius by one or more molecular markers. We've begun characterizing several hundred of these via a variety of classical microbiology and molecular biology techniques, and have already identified many unreported alleles and strains. A subset of this variation from MI was selected for MLST analysis (N=96 isolates sequenced for four loci). From these 96, 9 unique sequence types (ST) from two clonal complexes (CC3 and CC12) were identified, including 5 previously unknown ST in CC12. Combined with variation in mtxa plasmid positivity (a putative virulence gene on pMP19-plasmid), and several other molecular markers, we have several dozen unique strains ready for genomic sequencing and further study. To characterize their virulence, 30 of these M.plutonius isolates were tested in dose-response studies on in-vitro reared worker larvae. Most strains were highly virulent, but a few were apparently benign. We also confirmed bacterial inoculum concentration and survival in royal jelly and observed variable growth rates on agar media, broth, etc. These results suggest a veritable treasure trove of biological diversity among our collections, and we expect significant insights from their genomic sequencing and analysis, slated for Spring 2024. In addition, we've now completed all lab work and preliminary analyses on the N=700 larvae from our in-vitro PAMPs study reported last year. In brief: 24-hour-old larvae were assigned one of 8 treatments (2 controls plus 6 diets composed of either: gram-positive bacteria (M. plutonius), gram-negative bacteria (Serratia sp.), or one of four PAMPs (pathogen-associated molecular patterns) derived from gram-positive bacteria, gram-negative bacteria, and fungi. These are all known to trigger immune and oxidative stress responses in other animal systems, so we sampled larvae every 24 hours and measured their gene expression via qPCR for 13 markers associated with innate immunity and stress in honey bees. These results will be published in 2024. Objective 4: All of the work detailed above contributes directly and indirectly to our knowledge, detection, diagnosis and even potential treatments for honey bee disease. Specific contributions to diagnosis include both our optimized, pooled-sample workflows for molecular diagnostics and the complementary AI-image-diagnosis tool. Our pooled-sample qPCR approach has been adapted from methods introduced for mass COVID-19 qPCR testing. This has allowed us to screen hundreds of individual samples in pools (each composed of biologically similar replicates, e.g., 3rd instar healthy larvae), and then target our individual-sample qPCR on the groups and markers where pooled signals were significant. These molecular diagnoses are paired with a database that now includes >4,000 photos of diseased and healthy brood. These molecularly-diagnosed photos create an extremely high-quality training set, which is the basis for our rapid advances in creating an AI-based application to diagnose honey bee brood diseases via smart-phone photography. This BDBD-AI application will allow us to quickly and accurately identify different diseases based on visual characteristics of the larvae. By using machine learning techniques, we have built a working model on a small subset of the data and confirmed its ability to distinguish EFB from PMS-associated viral symptomologies. As detailed above, misdiagnosis of EFB and EFB-like PMS is a widespread problem observed in our data and results in significant costs to honey bees, beekeepers, and agriculture as a whole. The application will also be able to make recommendations for treatment based on the identified disease. This will be a valuable tool for beekeepers, researchers, and other professionals working with bees, as it will enable them to quickly and accurately diagnose and treat diseases in their hives. In 2024 we will focus on increasing the training set for this tool by working with other state apiary inspectors to provide us with non-EFB disease states, including Varroosis or Parasitic Mite Syndrome (PMS) and Idiopathic Brood Disease Syndrome (IBDS). We have finalized the details of a MTRA with Mann Lake Ltd. the major commercial interest that caters to beekeepers for the development and distribution of our AI diagnosis tool.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
Anderson, KE, Copeland, DC, Erickson, RJ, Floyd AS, Maes PW, Mott BM. 2023. A high-throughput sequencing survey characterizing European foulbrood disease and Varroosis in honey bees. Scientific Reports https://doi.org/10.1038/s41598-023-28085-2A
|
Progress 01/01/22 to 12/31/22
Outputs
Target Audience:Target audiences reached in 2022 include theAmerican Honey Producers Association - Annual Conference and Trade Showin Tucson, AZ, in December 2022. We have been invited to present our findings to theApiary Inspectors of America held in conjunction with the American Beekeeping Federationconcerning recent progressandsampling details needed to further pursue Image-Based-Diagnostic-Services for honey bee brood disease. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?A Post-doctoral student has been hired and trained in honey bee microbiology and larval disease including all classsical microbiology and molecular biology associated with diagnosis and genotyping. A PhD gradutate student isbeing trainded in the methods ofmulti-locus sequencing typing, and various undergraduate students have been trained in microbiology, in vitro larval rearing and the methods of molecular biology. Our collaboration with the apiary inspectors also involves professional development as much of this work serves to retrainthe human gestaltdiagnosis ofbrood disease with reference to image based diagnosis via machine learning methods. How have the results been disseminated to communities of interest?Results detailing our progress on brood disease diagnostics were presented at theAmerican Honey Producers Association - Annual Conference and Trade Showin Tucson, AZ, in December 2022. We have been invited to present our findings to the Apiary Inspectors of America concerning recent progressandsampling details needed to further pursue Image-Based-Diagnostic-Services for honey bee brood disease. We have furthered ourworking relationship with Dalan Health Inc. concerning the function and development oftransgenerational immune priming (TGIP) in honey bees, testing various EFB strains that differ in virulence and geographic prevalance. What do you plan to do during the next reporting period to accomplish the goals?Objectives 1 and 4) Determine the microbiome associated with larval health and disease, and Improve the detection and diagnosis of larval disease, both known and discovered.Through our collaboration with the Apiary Inspectors of America, we are collecting a wide variety of brood disease images and corresponding disease microbiomes. We lack the needed depth of information (images) from non-EFB disease states, including Varrosis or Parasitic Mite Syndrome (PMS) and Idiopathic Brood Disease Syndrome (IBDS).To provide the variation needed to increase the confidence of our predictive tool, wewill expand our AI-training set to include greater representation of non-EFB larval disease symptomology from throughout the US. Based on past results, we calculate that ten more sampling events of this size and character are needed to complete the training set for our digital tool. We will recruit approximately ten state apiary inspectors to take digital images of brood disease phenotypes and ship diseased brood samples directly to our lab for molecular diagnostic screening and subsequent microbiome analysis. We are applyingmachine learning to predict diseased microbiomes based on digital imagery, and our collaboration with the apiary inspectors provides the detailed images to further reinforce our training set. The postdoctoral researcher ispresently developing a Convolutional Neural Network (CNN) with SCINet resources to extract critical features from high-resolution digital images of larval disease phenotypes that possess corresponding microbiome data. CNNs are flexible and have already been successfully implemented to solve many image-based challenges in honey bees and other organisms.Because different brood diseases are associated with distinct symptomology, our present CNN can identify disease agents with a high degree of accuracy following training with a set of informed examples. Ultimately, this work will provide a new tool for beekeepers and apiary inspectors to quickly access image based diagnostic services. Objective 2)We will continuethe in vitrotesting focused on a variety of non-EFB opportunists identified in the microbiome screen. As an added dimension, we are collaborating with Dalan Health Inc. to understand the inheritance of EFB molecular patterns from queen to larvae and its effect on larval immunity. Objective 3)We continue todetermine the distribution of Melissococcus clonal groups and virulence genes within apiaries across landscapes, and throughout the US. We have amassed the DNA primer sets and geographic samples sets to begin thismonumental task, and lab work has begun, guided by a molecular screening for virulence.
Impacts
What was accomplished under these goals?
General:We've made significant progress on all four objectives in FY22.We filled our National Program awarded post-doc in October 2022, and this has greatly accelerated ourbrood disease research, including the development of AI-based diagnostic methods.In collaboration with the lab of co-PD Meghan Millbrath, we continued to sample larvae from multiple apiaries expressing various forms of brood disease throughout the state of Michigan, and now have three years of data from Michigan bluberries, a crop that results in predictable brood disease symptoms year after year. We have solidified our materials transfer with co-PD Jay Evans and the Beltsville lab, greatly improving our ability to study brood disease variation. Our relationships with the scientific and diagnostic communities have also expanded,facilitating our progress. Objective1:To date,we have sampled 31 different operations from 26 states consisting of 1151 total larvae resulting in 409 Bacterial isolates, and 1011 larvae with associated images. Sent from the Beltsville diagnostic lab in 2022, we have processed 107 samples from 24 states, comprised of 33 comb samples and 74 smear samples.From this collection,we have processed and sanger sequenced 107 isolates from 3 media types. Incollaboration with the Milbrath lab at MichiganState University, new samples were from 7 apiariesand 5 different operations resulting in a total of 911 individual larval samples, 140 fresh and 771 frozen. All of these samples have associated photographs, contributing needed variation to our AI training set.From Arizona,we sampled three apiaries showing brood disease symptoms, all of which were photographed in detail.This resulted in >100 samples corresponding to the melty phenotype (EFB-like disease) reinforcing our training set. Objective 2:We collected over 700 samples to assess larval immunity and oxidative stress gene expression through in vitro experiments. We fed 24-hour-old larvae a pathogenic strain of gram-positive bacteria (M. plutonius) and a gram-negative bacteria (Serratia sp.) isolated from diseased larvae. We also fed pathogen-associated molecular patterns (PAMPs) representing gram-positive bacteria, gram-negative bacteria, and fungal molecular patterns, which often trigger immune and oxidative stress responses in other animal systems. We performed RNA extractions on all samples and qPCR gene expression on roughly half of the samples for 13 gene markers. Gene expression was chosen to cover a range of processes associated with innate immunity including; antimicrobial peptide expression, oxidative stress, general metabolism, Toll, and IMD pathways: Abaecin, Catalase, CuZnSOD, Defensin1, DSCAM, Glutathione-S-Transferase1, Hymenoptacein, Insulin-like-peptide1, IMD Lysozyme1, PPO, TOLLand DUOX. The results are forthcoming in FY23. Objective 3:We have made substantial progress on Objective 3, isolating, cataloguing and collating samples acquired from the Beltsville diagnostic lab.We hadmany fruitful meetings with co-director Dr. Jay Evans and his collaborators to provide detailed processing of brood disease samples sent to Beltsville for diagnostic analysis. This material transfer has become streamlined, facilitatingour need for both fresh and frozen samples. Beginning this season, we have in place the infrastructure and molecular methods for the efficient processing of all samples sent through Beltsville for diagnosis. We previously reported a collaboration with an Illinois apiary inspector who provided us with a collection of larval stages and corresponding high-resolution photographs. We sequenced the 16S rRNA gene and interpreted the microbiome data, and quantifiedviral loads with qPCR and plasmid standards. Currentlyin review, these results will be published in early FY23, providingthe largest larval microbiome dataset to date, illustrating both health and disease, and totaling more than 1000 larval amplicon libraries. Using this initial training set, we began developing an AI-based application to diagnose honey bee brood diseases with visual imagery. This application will allow us to quickly and accurately identify different diseases based on visual characteristics of the larvae. By using machine learning techniques, we can train the application to recognize patterns and features that are indicative of specific diseases. The application will also be able to make recommendations for treatment based on the identified disease. This will be a valuable tool for beekeepers, researchers, and other professionals working with bees, as it will enable them to quickly and accurately diagnose and treat diseases in their hives. In FY23 we will focus on increasing the training set for this tool by working with other state apiary inspectors to provide us with non-EFB disease states, including Varroosis or Parasitic Mite Syndrome (PMS) and Idiopathic Brood Disease Syndrome (IBDS). Objective 4:We have thus far photographed and sampled diseased larvae identified in the field as EFB or EFB-like from 100+ apiaries. Next generation sequencing of the larval microbiome revealed various disease states including cryptic non-EFB brood diseases: IBDS, PMS, and chalkbrood mixed with EFB. Reflecting the value of the microbiome approach, we found a strong and significant relationship between symptomology and molecular results. The cryptic symptomology associated with the melty phenotype (also refered to as snot brood or melty brood) is difficult to distinguish from EFB based on human gestalt, but isrecognized with a high degree of confidence using detailed images and artificial intellegence.
Publications
|
Progress 01/01/21 to 12/31/21
Outputs
Target Audience:Communities served by the larval disease project include commercial beekeepers, hobbiests,and farmers. In addition, The PD is associated with an NIH educational grant, wherein this subject matter is discussed and disseminated to college students at hispanic serving institutions. Our approach included formal classroom instruction, laboratory instruction, and innovative teaching methodologies focused on experiential learning opportunities associated with agriculture and the honey bee microbiome. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?We continue to develop and nurture relationships with both hobbyist and professional beekeepers throughout the United States to correspond, teach and learn about larval disease, including sample metadataand photographic documentation of typical and atypical brood disease. Through our association with an NIH educational grant, this subject matter is discussed and disseminated to college students at hispanic serving institutions. Our approach included formal classroom instruction, laboratory instruction, and innovative teaching methodologies focused on experiential learning opportunities associated with agriculture and the honey bee microbiome. How have the results been disseminated to communities of interest?Results are mostly forthcoming. We have communicated with Scientific beekeeping.com, and the California beekeepers via phone,videoconferenceing and email. What do you plan to do during the next reporting period to accomplish the goals?An administrator awarded post-doc begins in early 2022, a factor that will greatly accelerate our progress towards understanding the larval microbiome in health and disease. We will continue to develop and nurture relationships with beekeepers throughout the United States to correspond about larval disease, including samples and photographic documentation of atypical brood disease. Over the next year we will refineprotocols, including a qPCR test forM. plutonius,and primer sets to screen for viral agents and opportunistic microbes. This protocol will save time and money by reducing the information and samples that proceed to the next step of the diagnostics and discovery protocol.
Impacts
What was accomplished under these goals?
Progress in general: The extended Covid-19 pandemic, including lab closures, precautions, maximized elework posture, and supply shortages continued to complicate our research efforts throughout FY20 and FY21. Other external factors of a local nature further delayed our work (e.g., building renovations, seasonal constraints, facility disruptions, etc.). Maximized telework and the loss of key personnel put an abrupt stop to planned in vitro lab tests. Our lab resumed more regular on-site activity in May-June of 2021, and we resumed training and coordinating lab work of three graduate students, one newly hired technician, and two newly-hired student aids. An administrator awarded post-doc begins in early 2022, a factor that will greatly accelerate our progress towards understanding the larval microbiome in health and disease. Progress on Objective 1: In collaboration with the lab of co-PD Meghan Millbrath, we continue to sample larvae from multiple apiaries expressing various forms of brood disease throughout the state of Michigan. As hives moved throughout various pollination routes, we recorded shifting disease states collected by longitudinal and replicated sampling, including covariate hive metrics from both recovered and deceased hives. Since May of 2021, we have processed many isolates in the microbiology lab to characterize the bacteria associated with various disease states. We have also collated the most curated and informationally rich sample sets for next generation sequencing effort. We continue to develop and nurture relationships with beekeepers throughout the United States to correspond about larval disease, including samples and photographic documentation of atypical brood disease. Some of these collections are presently prepped for deep sequencing following a molecular diagnostic protocol that deductively eliminates known causes of brood disease. Concurrently, from the state of Illinois, we continued our metagenomic exploration of previously collected disease states and larval stages as designated by high-resolution photographs and extension professionals. We prepared, sequenced, and analyzed the 16S gene from 850 additional larval disease microbiomes from 18 apiaries in Illinois, bringing our grand sequencing total to 1070 larval microbiomes from 24 Illinois apiaries. We increased the size of our microbial culture collection, and genome sequenced a number of disease related pathogen strains to complement our planned laboratory work. We established a research collaboration to explore the development of immunity throughout larval stages and early adult growth. We will continue to explore these factors in relationship to the larval microbiome, and the introduction of molecular patterns in the queen diet. We are currently analyzing large sets of larval amplicon data to detect various patterns associated with health and disease. We will explore this phenomenon more comprehensively this coming season. Progress on Objective 2: Following the loss of key personnel, we retrained an individual to perform in vitro larval rearing, producing results on environmental context, and host susceptibility relative to colony context. Working with commercial beekeeping operations in Michigan, we received larval samples representing idiopathic brood disease. We extracted DNA and RNA from these samples and are currently analyzing BactQuant yields and immune gene expression to compare with the disease phenotype descriptions. Based on diseased larvae samples collected from throughout the US, we recovered a variety of M. plutonius and other disease-associated bacteria designated as idiopathic. According to Sanger sequencing of the total 16S rRNA gene, we recovered over 300 isolates via culture-based methods with 65% of these isolates representing Melissococcus plutonius strains. To understand microbe interactions with honey bee antimicrobials, we tested a taxonomic variety of disease associated strains against various concentrations of royal jelly and honey in vitro. Using negative controls isolated from non-Apis environments, Staphylococcus epidermidis and Escherichia coli, we tested the ability of various microbes to grow in concentrated and dilute honey and royal jelly. Controls performed as expected validating the assay. In general, royal jelly was more inhibitory than honey. We found that most of the disease associated strains isolated from sick larvae could proliferate in concentrated honey and royal jelly, suggesting they are well evolved to exploit this environment. Gammaproteobacteria (Gilliamella spp. and Frishella spp.) either related to, or classified as core gut bacteria of adults were inhibited to varying degrees in both honey and royal jelly, suggesting they rely on To understand larval immunity and oxidative stress gene expression, we setup different in vitro trials feeding young larvae bacteria and pathogen associated molecular patterns (PAMPs). We used a pathogenic strain of M. plutonius and a Serratia strain cultured from diseased larvae as comparative controls. To represent the response to PAMPs, we chose a variety of molecular patterns including mannan, lipopolysaccharide, peptidoglycan, and β 1-3glucan. We recorded survival and surviving larvae were used in downstream analysis of gene expression. We assayed gene expression to evaluate factors that promote larval survival. These results will help differentiate cause from consequence and suggest approaches and compounds to mitigate disease, and reduce colony loss. Progress on Objective 3: We have researched, designed and tested multiple primer sets to perform multi-locus sequence typing (MLST) analysis of European foul brood disease in the United States. We had three meetings with co-director Dr. Jay Evans and his collaborators directed towards more direct and detailed processing of brood disease samples sent to Beltsville for diagnostic analysis. This material transfer has proven difficult due to the present situation with the pandemic closures, laboratory workspace and time and our need for both fresh and frozen samples. Beginning this season we have in place the infrastructure and molecular methods for the efficient processing of all samples of brood disease sent to (through) Beltsville for diagnosis. We have also formalized methods for partitioning samples of putative larval disease sent for diagnostic services to the Beltsville MD lab. Progress on Objective 4: We developed and implemented a protocol to diagnose unknown brood disease samples, including qPCR of bacterial and fungal abundance, and species-specific molecular tests (presence absence PCR) for Paenibacillus larvae and Melissococcus plutonius, the causative agents of American and European foulbrood respectively, and sacbrood virus. We will perfect these protocols in the new year, including a qPCR test for M. plutonius, and primer sets to screen for viral agents and opportunistic microbes. As an application of principle, we applied this protocol to diagnose an unknown larval disease shipped from the island of Molokai in Hawaii, May 2021. The outbreak was feared to be American foul brood, the most devastating of honey bee diseases. Our results revealed only two of 48 samples tested positive for Paenibacillus larvae (AFB), while nearly all samples contained an abundance of Melissococcus plutonius EFB. Nucleic acid sequences corresponding to sacbrood virus were not detected. This protocol saves time and money by reducing the information and samples that proceed to the next step of the diagnostics and discovery protocol.
Publications
- Type:
Journal Articles
Status:
Submitted
Year Published:
2022
Citation:
Anderson KE, Floyd AS, Mott BM, Maes PW. 2022. Microbiomes associated with European foul brood disease, and idiopathic brood disease syndrome in honey bees. Insects. submitted
|
|