HoneyMicroAtlas: Comparative Analysis of Worldwide Honey Microbial Fingerprints and Their Ecological Roles

1. Overview

The honey microbiota represents a structured and functionally relevant microbial ecosystem, typically dominated by Lactobacillaceae bacteria. Apis Honey is the most consumed and has been extensively studied with metagenomic resources. On the other hand, honey produced by stingless bees of the tribe Meliponini is widely recognized for its medicinal properties^1,2. These bees are cultivated by Indigenous communities in Latin America and hold deep biocultural significance^3,4. In 2025, we participated in identifying “Meliponilactobacillus,” a bacterial species closely related to Acetilactobacillus and Nicoliella that is consistently present in the honey microbiome of Meliponini worldwide⁵. Our goal is to elucidate the ecological roles of the microorganisms that define the core microbial fingerprints of honey, and, as proof-of-concept of their ecological relevance to investigate their interactions with stingless bee hosts. For these reasons, we propose integrating multiple datasets and reanalyzing publicly available honey metagenomes from both Apis and Melipona species into an interactive comparative dashboard. This platform will enable taxonomical and functional comparison, generating microbial “fingerprints” of honey types across geographic regions. In parallel, to better understand the ecological role of dominant Lactobacillaceae lineages, we will generate targeted complementary data to explore its influence in the bee host.

Using bees as a model provides a simpler system for studying the microbiota–gut–brain axis. Unlike vertebrates, whose complex gut microbiota makes it difficult to distinguish bacterial metabolites from host-derived metabolites, insects harbor simpler microbial communities, making them valuable models for investigating the evolutionary mechanisms through which microbes influence brain metabolism^6,7. We have direct access to meliponaries in Quintana Roo (Mexico) and Costa Rica, where honey is sustainably harvested. The meliponaries have formally endorsed our research goals. Laboratory work will be conducted in the Molecular Biology Laboratory, while computational analyses will be performed in the Computational Biology Laboratory of the Centro de Ciencias Matemáticas, which hosts two high-capacity servers (750 GB RAM each).

Once we receive the funding, we estimate 3–4 months will be required to generate the data and 3 months for analysis. We anticipate completing the project within 6 months. The approximate timeline is as follows: On the bioinformatic side Month 1, download honey microbiomes and quality control, Month 2-3 Diversity analysis. Month 4 classification algorithms and microbial fingerprints calculation. Month 5-6 platform development. On the experimental side we planned months 1-2 Microbial cultivation, extraction-sequencing, and Inoculation experiments. Months 3-5 Metabolomics, and Months 5 and 6 Data analysis.

All sequencing data will be deposited in the NCBI Sequence Read Archive (SRA), and all analysis scripts and processed results will be made publicly available through appropriate repositories (GitHub/Zenodo) and a public dashboard.

2. Introduction

Recent studies show that honey origin can be inferred from its microbial composition⁸, underscoring the relevance of honey microbiomes as traceable food-associated ecosystems. Just as molecular fingerprints have been established for urban microbiomes^9,10 through large-scale data integration, we are now in a position to approach honeys from a similar perspective. We will develop a comparative, publicly accessible dashboard to analyze honey metagenomes worldwide. Based on our previous work⁵, we expect Meliponilactobacillus, closely related to Acetilactobacillus and Nicoliella, to be a defining component of the molecular fingerprint of Melipona honeys. Our previous analyses of 17 mature pot-honey samples from two native stingless bee species, Scaptotrigona mexicana and Melipona beecheii, consistently detected a dominant Lactobacillus signal and yielded multiple MAGs representing an Acetilactobacillus-like lineage⁵. These MAGs shared ≤80% ANI with public Acetilactobacillus genomes, indicating previously uncharacterized diversity and supporting a novel taxon we provisionally designate Meliponilactobacillus. We also obtained six isolates from the same material, most of which clustered closely with our MAGs. Given their prevalence and abundance, these bacteria may contribute to shaping the honey microbiome and to its bioactive properties, including reported antioxidant and antimicrobial activities.

Based on these findings, we now aim to investigate when and how Meliponilactobacillus becomes abundant during honey maturation and its functional role in stingless bees. Previous studies in Apis and Bombus have shown that specific microorganisms can modulate metabolite abundance and gene expression in the brain, but equivalent research is still lacking for stingless bees. This absence is largely due to the previous unavailability of axenic models, which are essential for testing microbial colonization under controlled conditions. Preliminary experiments conducted at the Laboratory of Microbial Ecology and Genetics (Universidad de Costa Rica) and at CINAT have demonstrated the feasibility of rearing axenic stingless bees. As a next step, we aim to initiate controlled inoculation experiments with selected microorganisms, such as Meliponilactobacillus, followed by brain metabolomics analyses, to explore the possible mechanism underlying this symbiosis.

Carrying out inoculation with Melliponilactobacillus is of particular interest, since members of Lactobacillus have been shown to influence brain metabolite abundance and gene expression associated with behaviour in the other two social groups of corbiculate bees, honeybees and bumblebees^11,12. We also aim to investigate whether the honey microbiome changes over time, for example, whether the concentrations of these bacteria vary depending on honey maturation time, comparing freshly sealed pots with honey collected by beekeepers. This grant would help not one but a team of female authors, 2 students, and three early career researchers to finish this project in which we are all interested.

In synthesis, we want to test the following hypothesis:

H1 Molecular fingerprints allow to differentiate honey by its origin (comparative visualization in a dashboard)

H2 Honey maturation and influences the microbiome composition in time, including the distribution of Meliponilactobacillus.

H3Brain metabolite abundance changes when inoculating the bees with Meliponilactobacillus.

Impact statement

Honey is the bees’ food during winter and part of the larvae’s diet; this study will lay the groundwork for understanding whether this microorganism contributes to the stability and resilience of honey as a preserved food source. The results of the experiments testing these hypotheses will enable us to enhance our understanding of the role of Meliponilactobacillus in relation to other microorganisms and to the bees themselves. If Meliponilactobacillus starts in low abundance in the nectar, and then dominates the microbiome after honey maturation, then it may play a role in defending honey against other microorganisms or it may have a role in bees maturation. By linking microbial dynamics to brain metabolite changes, we will explore potential gut–brain interactions in bees, opening new avenues for stingless bees health research. Together, these findings could have significant implications for pollinator conservation, honey production, and our broader understanding of host–microbe interactions.

3. Methods

Participants/Sample

Reanalyzing data. We have collected public Apis mellifera honey microbiome metabarcoding and metagenomic data (n~120) that, together with our Mellipona honey microbiomes, (n=17) will be reanalyzed to search for taxonomic assignation, MAGs and biosynthetic Gene clusters.

Sampling: Honey samples will be collected from pots produced by stingless bees (e.g., Melipona beecheii, Scaptotrigona mexicana) sourced from Yucatán Península, México, and Costa Rica. Each pot constitutes a sampling unit for metabarcoding profiling of the honey-associated microbiome. To follow microbiome dynamics, we will collect two samples (biological replicates) at three time points: open pots with nectar (t=0), freshly sealed honey pots (t=1), and mature honey (t=2). To investigate host-interaction, we will grow our isolated Meliponilactobacillus from mature honey. For the inoculation experiments, mature pupae will be extracted from meliponaries located at CINAT (Heredia, Costa Rica). Rearing of bees: At least 48 mature pupae will be collected and reared in sterile boxes. Pre-inoculation: Emerged bees are verified to be microbiota-depleted by qPCR before inoculation. Inoculation: Bees are exposed to Melliponilactobacillus. Post-inoculation: metabolomics analysis from at least 12 bees exposed to the inoculum and 12 bees from the control treatment.

Inclusion criteria: We will select nectar/honey samples that are free of visible contamination at the time of collection. For the inoculation experiment, all the bees that successfully develop and come from boxes where the microbiota depletion test by qPCR shows no detectable bacterial DNA.

Exclusion criteria: Evidence of environmental contamination. Insufficient material for DNA extraction. Samples with few reads after filtering or of poor quality. For the inoculation experiment, boxes that had a bee with amplification in the qPCR.

Sampling strategy: Based on previous shotgun observations of mature honey, we believe that Meliponilactobacillus will dominate the microbiome regardless of bee species and location. Therefore, we will stratify sampling in collaboration with local producers, targeting representation across bee species, locations, and time points. Manual honey extraction will be performed on one to three hives, and the honey will be distributed into eight sterile, pyrogen-free 50mL tubes. This is to ensure the quantity and quality of DNA during the extraction. Metadata will be collected, including bee species, location (city and coordinates), collection date, humidity, pH, sugars, storage conditions, and pot identifier.

Sample size justification: Priority will be given to (i) covering several regions/meliponaries of the Yucatán Peninsula and (ii) maintaining the three temporal samplings per pot to compare microbiome changes within each pot over time. Preference will be placed on achieving sufficient sequencing depth and expanding geographic coverage rather than increasing the number of pots from the same site. We will collect four biological replicas (different bee species and location), even though we only sequenced two to adjust the budget. Given that we expect Meliponilactobacillus to be dominant in mature honey, two biological replicates at three time points would be sufficient to demonstrate this microbial dynamics. With a small dataset, it will still be possible to assess temporal dynamics and contrast patterns across regions; the analyses (Bray–Curtis, NMDS, PERMANOVA with pot as strata) are consistent with a reduced N.

A total of 48 pupae will be used in the experiment, equally divided between the control (24 pupae) and inoculated (24 pupae) groups. Potential mortality is taken into account to ensure that at least 12 adult bees per group will be available for metabolomic analyses. Although the sample is small, it is statistically sufficient to detect biologically relevant differences between groups⁹, while also optimizing the use of resources and ensuring feasibility within the scope of the project.

Materials and Procedures

DGenomic DNA Extraction and Purification: We will follow the DNA extraction workflow previously optimized in our group and collaborators⁵. Sample input: DNA will be extracted and concentrated from 50 g of honey into 1.5 mL tubes. Pre-treatment: A pre-treatment step will be applied following protocols in Soares 2011 ¹⁶, and Waiblinger 2012¹⁷ to optimize DNA recovery from honey matrices. Extraction: DNA will be isolated using the Qiagen DNeasy Blood & Tissue Kit after determining the optimal dilution for each sample. Mechanical lysis will be performed with a TissueLyser II (Qiagen) using 3 mm tungsten carbide beads. Column-based purification: Silica spin columns will be used following manufacturer’s guidelines (buffer ATL, Proteinase K, ethanol) to obtain DNA of sufficient yield and purity. Quality control: DNA concentration will be quantified (≥20 ng/µL, total ≥500 ng), and purity assessed with a QIAxpert spectrophotometer, targeting A260/280 ratios>1.7. Microorganism sampling: Complementary microbial isolation will be conducted following Syed 2022¹⁸ methodology

Bee rearing: Anexic and gnotobiotic bees will be produced following the methodology previously used in honeybees and bumblebees ^19,20. Briefly, mature pupae will be collected from meliponaries at CINAT, placed in sterile boxes, and incubated to complete their development. Pupae kept under sterile conditions will emerge as microbiota-free adult workers. At last, six boxes with eight pupae will be reared, resulting in a total of 48 bees. To confirm that bees are microbiota-free, the gut from one bee per cage will be extracted, and bacterial loads will be determined by quantitative PCR (qPCR) using universal primers targeting the 16S rRNA gene. If a box shows amplification in the aPCR assay of its pupae, it will be excluded from the experiment. Inoculation: Bees that have emerged with no development defects will be selected and transferred to new sterile cup cages placed on top of 100 mm petri dishes, where they will be exposed to the inoculum. The inoculum (Melliponilactobacillus) will be mixed with sterile sugar water (1:1), and added as 100 ul droplets. Control cages will contain bees that will be fed only with sugar water. From the total number of successfully developed bees, half will be inoculated, and the other half will serve as a control. The exposure to the inoculum will be for 7 days. Metabolomics: Brains will be dissected for at least 12 bees from each group for metabolomics analysis. Samples will be processed using a modified Bligh and Dyer protocol, and a GC-MS analysis of metabolites will be performed ²¹.

Planned Analyses

Metabarcoding (amplicon metagenomics)

We will follow the typical metabarcoding analysis pipeline with Chime, starting from quality control to diversity assessment. The main steps include quality control, trimming and filtering, and ASV taxonomic assignment using GreenGenes as the reference database. We will complement the reference database by incorporating Meliponilactobacillus 16S genes with the new MAGs reported by Xolalpa et al. 2025⁵. For the taxonomic and diversity analyses, we will assess the honey microbiome using classical metagenomic diversity indices using the phyloseq package in R and scikit-bio library in Python: Beta diversity: Bray–Curtis distances with NMDS visualization, Alpha diversity: richness and diversity indices to evaluate community structure across samples and abundance plots: to compare microbial composition across conditions.

Differential abundance between nectar, freshly sealed honey, and harvested honey will be assessed using methods such as PERMANOVA (for beta diversity) and DESeq2 or Wilcoxon rank-sum tests (for specific taxa), to determine whether observed differences are statistically significant. We will also explore the integration of public Apis mellifera honey data to utilize supervised machine learning methods, such as random forests and support vector machines, to predict whether honey samples originate from freshly sealed pots or from mature honey.

Metabolomics:

Best hits of compound identity will be reported for spectra with a match factor >85%. Identified metabolites will then be manually mapped to metabolic pathways in the KEGG PATHWAY Database. Analyte abundances will be calculated using the MassHunter Workstation Quantitative Analysis software (Agilent). The impact of colonization on metabolite abundance will be calculated using a mixed integer linear model, as implemented in the lmm2met package in R.

Ethics and Data Management

Ethical approval will be obtained from the relevant institutional bodies when required. Confidentiality will be ensured by anonymizing sensitive information. All generated data will be shared in accordance with FAIR principles and deposited in publicly accessible repositories, while sensitive metadata will be archived under controlled-access conditions.

Pilot Data: We already have 17 honey samples analyzed using shotgun metagenomics, which will serve as a point of comparison. Since Meliponilactobacillus is a new species to perform taxonomical classification, we need to add 16S sequences as references to the Silva/Green Gene databases and the whole Genome to krakenDB. To add these references, we will extract the 16S sequence from the 25 MAGs obtained from these 17 samples. In addition, at the Centro de Innovación para el Desarrollo Apícola Sustentable en Quintana Roo (CIDASQROO) of the Universidad Intercultural Maya de Quintana Roo, samples of Melipona beecheii and fermented honey are available; in the event of difficulties, we would not have any difficulty obtaining new samples.

We implemented the protocol for producing axenic and gnotobiotic bees, previously used in Apis and bumblebees, and confirmed that it is also effective for generating microbiota-free stingless bees suitable for subsequent inoculations.

5. Budget

Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data.

Sampling (covered by us)
Amplicon metabarcoding sequencing: 6 samples 1500 DOL
Student scholarship 5 months 1000 DOL
Metabolomics (24 samples 2500 DOL)

6. References

1. Cabezas-Mera, F., Cedeño-Pinargote, A. C., Tejera, E., Álvarez-Suarez, J. M. & Machado, A. Antimicrobial activity of stingless bee honey (Tribe: Meliponini) on clinical and foodborne pathogens: A systematic review and meta-analysis. Food Front. 5, 964–993 (2024).

2. Ramón-Sierra, J., Magaña-Ortiz, D., Ortiz, E. & Ruiz Ruiz, J. Medicinal uses of Melipona Bee Honey in Mayan Communities of Yucatán. in Stingless Bee’s Honey from Yucatan: Culture, Traditional Uses and Nutraceutical Potential (eds Ortiz Vázquez, E. de la L., Ruiz Ruiz, J. C., Magaña Ortiz, D. I. & Ramon Sierra, J. M.) 11–26 (Nova Science publishers, 2016).

3. Ek, J. & Magaña, M. Características socioeconómicas de la meliponicultura en localidades rurales del estado de Yucatán. Textual 205–232 (2023) doi:10.5154/r.textual.2022.81.11.

4. Lazarino, L. da C. et al. Meliponicultura: potencialidades e limitações para a conservação de abelhas nativas e redução da pobreza. Divers. J. 6, 2217–2236 (2021).

5. Xolalpa-Aroche, A. et al. Uncharacterized Acetilactobacillus-like genera represent the most abundant microorganisms in honeys produced by native stingless bee. 2025.06.25.661387 Preprint at https://doi.org/10.1101/2025.06.25.661387 (2025).

6. Ashique, S. et al. Gut-brain axis: A cutting-edge approach to target neurological disorders and potential synbiotic application. Heliyon 10, e34092 (2024).

7. Engel, P. & Moran, N. A. The gut microbiota of insects – diversity in structure and function. FEMS Microbiol. Rev. 37, 699–735 (2013).

8. Sabater, C., Calvete, I., Vázquez, X., Ruiz, L. & Margolles, A. Tracing the origin and authenticity of Spanish PDO honey using metagenomics and machine learning. Int. J. Food Microbiol. 421, 110789 (2024).

9. Danko, D. et al. A global metagenomic map of urban microbiomes and antimicrobial resistance. Cell 184, 3376-3393.e17 (2021).

10. Contreras-Peruyero, H. et al. CAMDA 2023: Finding patterns in urban microbiomes. Front. Genet. 15, (2024).

11. Liberti, J. et al. The gut microbiota affects the social network of honeybees. Nat. Ecol. Evol. 6, 1471–1479 (2022).

12. Li, L. et al. Gut microbiome drives individual memory variation in bumblebees. Nat. Commun. 12, 6588 (2021).

13. Wu, J. et al. Metagenomics and Metagenome-Assembled Genomes: Analysis of Cupei from Sichuan Baoning Vinegar, One of the Four Traditional Renowned Vinegars in China. Foods 14, 398 (2025).

14. Li, Q. et al. Insight into the acid tolerance mechanism of Acetilactobacillus jinshanensis subsp. aerogenes Z-1. Front. Microbiol. 14, (2023).

15. Chen, L. et al. Deciphering the acidophilia and acid resistance in Acetilactobacillus jinshanensis dominating baijiu fermentation through multi-omics analysis. Food Microbiol. 125, 104655 (2025).

16. Soares, S., Amaral, J. S., Oliveira, M. B. P. P. & Mafra, I. Improving DNA isolation from honey for the botanical origin identification. Food Control 48, 130–136 (2015).

17. Waiblinger, H.-U. et al. In-house and interlaboratory validation of a method for the extraction of DNA from pollen in honey. J. Für Verbraucherschutz Leb. 7, 243–254 (2012).

18. Syed Yaacob, S. N., Huyop, F., Misson, M., Wahab, R. A. & Huda, N. Exploring the genome of Lactobacillaceae spp. Sy-1 isolated from Heterotrigona itama honey. PeerJ 10, e13053 (2022).

19. Bonilla-Rosso, G. & Engel, P. Functional roles and metabolic niches in the honey bee gut microbiota. Curr. Opin. Microbiol. 43, 69–76 (2018).

20. Zheng, H., Steele, M. I., Leonard, S. P., Motta, E. V. S. & Moran, N. A. Honey bees as models for gut microbiota research. Lab Anim. 47, 317–325 (2018).

21. Gonzalez-Riano, C., Garcia, A. & Barbas, C. Metabolomics studies in brain tissue: A review. J. Pharm. Biomed. Anal. 130, 141–168 (2016). List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).

HoneyMicroAtlas: Comparative Analysis of Worldwide Honey Microbial Fingerprints and Their Ecological Roles

HoneyMicroAtlas: Comparative Analysis of Worldwide Honey Microbial Fingerprints and Their Ecological Roles

Funding Progress

Related Funding

Request for Proposals - Genomics and Proteomics

Nonprofit Recipient

Topics

DOI