https://orcid.org/0000-0002-6498-3275
Figures
Abstract
Efforts to improve honey bee colony health continue due to persistent high loss rates. A major focus in this area is Deformed wing virus (DWV), a key driver of colony loss. The application of modern molecular techniques has characterized the DWV genome and its high mutational rate that enables the formation of diverse quasi-species populations capable of evading host immune responses, while other work has led to the development of DWV clones suitable for sequence-specific tracking of viral dynamics. In this work we combine knowledge of these efforts to track the mutational progression in a DWV clone surrounding an area of low nucleotide diversity and compare it to its wild-type source. We achieve this through amplicon sequencing of the structural viral protein, VP2, after incubation across three generations and multiple host genetic sources. Inocula were injected into pupae, allowed to replicate, then extracted for a further two generations of injections. For the final injection generation, recipient pupae were injected with preparations from either the same genetic source or cross-fostered from other colonies. Overall, we compared the mean number and type of mutations, their proportional abundance in the read pool, and specific locations across strains. Sequencing results indicate a limited number of mutational hotspots, which were driven by silent mutations in the final injection generation of the wild-type strains. No significant differences were found among other mutation types, cross-fostering status, or interactions with host genetics. This work is an initial attempt at examining viral dynamics in a cloned system across multiple generations and treatment groups. The results provide valuable insights, which may further enhance our understanding of viral dynamics and potentially improve future honey bee therapeutics.
Citation: Nearman A, Buller-Jarrett A, Boncristiani D, Ryabov E, Chen Y, Evans JD (2025) Generational mutation patterns in a honey bee Deformed wing virus via infectious clones. PLoS One 20(11): e0337191. https://doi.org/10.1371/journal.pone.0337191
Editor: Olav Rueppell, University of Alberta, CANADA
Received: July 23, 2025; Accepted: November 5, 2025; Published: November 19, 2025
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All metadata and summary analysis from sequence data is available as supplemental information (S1 Table). All sequence data is available at NCBI under BioProject ID: PRJNA1295475.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Managed honey bee colonies in the US continue to experience high loss rates [1,2], due to a combination of factors, including pathogens, parasites, poor nutrition, and pesticide exposure – each acting alone or in synergy [3–6]. Among the various contributing factors, Deformed wing virus (DWV) is of particular concern due to its association with colony loss [7–9] and increased virulence when vectored by the ectoparasitic mite, Varroa destructor [7,10,11]. The need for new knowledge and potential solutions is great, as multiple strains of DWV have been detected worldwide [12–15].
Advances in next-generation sequencing technology have begun to reveal key insights into viral molecular physiology. Portions of the DWV genome display high polymorphic diversity, which may generate quasi-species capable of evading host immune responses [16]. On the other hand, identifying regions of low diversity may reveal immutable amino acid residues involved with cellular entry or other key sites for infectivity. Recent developments allow the successful creation and implementation of viral clones based on the DWV genome [17,18]. While these clones were motivated by efforts to encode reporter proteins for visualization, they can also be used to study mutational processes within a known population. This information could then be used to identify the effect of host variability on viral replication, thus increasing the resolution of previous studies characterizing variable host immune responses to pathogens [19–23] and studies of viral evolution under different modes of transmission [24]. Further, while RNA interference shows promise as an antiviral control strategy for bees [25,26] and indeed is a key endogenous immune response [27], studies exploiting cloned variants of DWV show that RNAi is less effective against infection when targeting regions with high sequence diversity [16]. Locating regions in the DWV genome that differ in mutation rate can lead to more effective controls that avoid hotspots for diversity.
Here we examined the distribution, abundance, and type of mutations accumulated by both a clone of DWV (NanoLuc) strain [18] and wild-type source populations (Wild Type) across multiple generations and within or between multiple colonies. We hypothesized that mutations would accumulate over successive generations and occur in similar genomic locations, regardless of viral source. Additionally, we anticipated that host genetic variation would influence mutation patterns. To test these hypotheses, we designed pairs of primers for clone-specific and non-specific amplicons encompassing a 433 nucleotide region of the gene encoding for the structural viral protein, VP2 (Fig 1). This region is directly adjacent to the clone insert, making it ideal for primers either spanning the insert and VP2 or the wild-type Leader Protein and VP2, thus selecting for each strain. We then produced amplicons before and after infecting hosts for three generations. The results suggest a generational accumulation of silent mutations in genomic locations that are similar across clone- and wild-type derived quasi-species, while no host effects were detected. This study represents the first attempt at examining location-specific mutations in a cloned honey bee virus under controlled experimental conditions.
RNA extractions were prepared and sequenced for each injection round as described.
Results
Recipient colony affects viral load, but there is no interaction effect. In consecutive trials viral load did not differ significantly among colony recipients in rounds 1 and 2 (Fig 2). Wild-type virus was present at higher copy numbers in both the initial inoculum and in recipients, as expected. Colony A showed higher viral loads than did Colonies B and C across all inocula in round 3, and there was no significant interaction effect whereby certain pairings resulted in unusually high or low levels of infection in round 3 (Supplemental S1 Fig).
Positional summary suggests region-specific accumulation of mutation types across injection generations. Regardless of strain, early sequence base variations (bases 150–200) are more prone to missense mutations than silent mutations but rarely accumulate above 1% proportional representation. By the final injection generation, silent mutations have site-biased accumulation in the wild-type strain at proportions greater than 1% representation. Overall, the wild-type strain indicates a greater accumulation of quasi-species, though similar site-specific patterns were observed in both strains (Fig 3).
Horizontal line identifies 1% proportional representation.
Wild-type strains accumulated a greater number of unique mutations driven by silent mutations. Across all samples, wild type strains accumulated a normalized average of 15.2 + /- 0.8 mutations compared to the NanoLuc cloned strain with 12.2 + /- 0.5 (W = 28, p < 0.004). This appears to be due to the mean number of silent mutations (wild type 7.4 + /- 1.0, NanoLuc 3.0 + /- 0.3, W = 4, p < 0.001), whereas they did not differ in their number of missense mutations (wild type 7.8 + /- 0.6, NanoLuc 9.1 + /- 0.5). Comparison of nonsense mutations was not possible due to low counts after normalization (Fig 4).
Significant differences by Wilcoxon Rank Sum Tests were observed between the global number of mutations between strains (W = 28, p < 0.004) and the number of silent mutations between strains (W = 4, p < 0.001). No comparisons between nonsense mutations were made due to representation in only two samples.
Wild-type strains showed a greater proportion of reads with silent mutations in the F2 generation. Similar to the number of unique mutations, their proportional representation in the read pool was also greater among wild-type strains (1.6e-2 + /- 2.6e-3) compared to NanoLuc strains (8.2e-3 + /- 1.3e-4, W = 0.5, p < 0.001). This appears due to the success of silent mutations (2.3e-2 + /- 4.4e-3, W = 2, p < 0.001) in the F2 generation (3.0e-2 + /- 5.1e-3, W = 0, p < 0.05). No significant differences were found related to the effects of cross-fostering, source colony, or target (end) colony (Fig 5).
Significant differences by Wilcoxon Rank Sum Tests were observed between the global number of reads with mutations between strains (W = 0.5, p < 0.001), the number of reads with silent mutations between strains (W = 4, p < 0.001), and the number of reads with silent mutations across generations for the wild type strain (W = 0, p < 0.05). No comparisons between nonsense mutations were made due to representation in only two samples.
Hierarchical clustering of normalized read counts by sequence position supports strain and generation-specific relationships among samples. Combining information on mutation location and proportional representation into a dendrogram reveals three major clusters of samples, separating F2 generation. T wild-type samples into two distinct clusters. The nonspecific sub-clustering among NanoLuc samples indicates little change in the mutational space but remains distinct from the F0 and F1 wild-type samples within the same cluster (Fig 6).
Each sample can be identified by the DWV strain type (NL = NanoLuc, wt = Wild Type), injection generation (F0, F1, F2), and second injection generation source colony (A, B, or C) or third injection generation colony cross-fostering status (denoted by “x”).
Reference-free calculation of Amplicon Sequence Variants (ASVs) reveals strong relationships across dominant quasi-species and overall counts. Raw counts of ASVs indicate a decrease in F1 followed by an increase in F2 on average for both strains when filtered to a proportional representation >0.1%. This relationship holds true for wild-type samples when filtered >1% representation (S2 Fig). Multisequence Alignment and maximum likelihood tree generation indicate that proportionally dominant quasi-species tend to be closely related, regardless of strain, generation, or cross-fostering status (S3–S5 Fig).
Discussion
This study represents a first attempt to explore viral sequence dynamics using a clone of the honey bee virus, Deformed wing virus (DWV). We achieved this by injecting a prepared inoculum through successive generations, either within or across multiple genetic lines, and measuring the mutational effects of nucleotide-level changes for the VP2 structural protein. The results suggest several region-specific mutational hotspots that persist across generations (Fig 3). The type of type, either silent or missense, remained constant within and across generations for the cloned virus. This was true for both the overall number of mutations (Fig 4) and their proportional representation in the read pool (Fig 5). Unsupervised clustering of both mutation location and normalized read count agree that there is no discernable effect observed for injection generation, cross-fostering, or host genetic influence on the mutational space (Fig 6). These results are further confirmed by reference-free quantification and maximum likelihood trees of amplicon sequence variants (S2–S5 Fig).
The presence of wild-type DWV in our samples enabled direct comparison of dynamics between the two strains. Notable is the significant accumulation of silent mutations by the final injection generation of the wild-type strain, in both overall count and proportional representation (Fig 4,5). These effects were also observed through unsupervised clustering (Fig 6), though the location of accumulated mutations does not appear to differ between strains (Fig 3). In terms of overall number of silent mutations, we did not observe a significant effect of injection generation, but a clear trend of accumulation of silent mutations can be seen between both strands (Fig 4). More injection generations may be required to accurately describe this trend, but the gradual accumulation may help explain the virus’ ability to evade immune responses, and RNAi in particular [25].
Significant differences between the NanoLuc and wild-type strains may be explained in part by the 10-fold difference in viral loads (Fig 2). Our qPCR results indicate greater replication rates of wild-type populations, perhaps due to slightly slower replication of NanoLuc clones thanks to added transcription or translation costs [28,29]. Initial development of similar cloned viruses displayed a low-level selection toward insert deletion, though overall stability remained intact at 72 hours post injection [17]. Further, clone-derived DWV Nanoluc [17,18] was based on the DWV variants isolated in 2015 [16]. It is possible that contemporary 2023 DWV variants are better adapted to the bees used in this study compared to 2015. Our qPCR results also reveal a slight decrease in viral loads by the final generation for both strains. This may indicate a decrease in primer efficacy for some quasi-species or an accumulation of non-viable strains in the final inoculum.
That the wild-type strains were present in our samples suggests two possibilities: 1) some level of wild type was present in the initial inoculum or, 2) a latent infection became active when the specimens (bee pupae) used for inoculum preparation were immune challenged [30–32]. The low number of mutations found in the prepared inoculum suggests the specimens were not experiencing an active infection at the time of injection. Also, the small number of closely related quasi species in the inoculum, relative to the final injection generation, supports this thinking (S3 Fig). Furthermore, rapid selection for insertions or deletions is improbable over these time periods, given the known excision rates [18].
The gradual accumulation of silent mutations across generations was unsurprising. The identification of the amino acid changes that lead to successful quasi-species, however, may elucidate further patterns in host-virus interactions. For example, 70% of samples contained a silent mutation at base position 346 due to third-position wobble, with just under half being a cytosine to thymine transversion (S1 Table). This information may prove useful for the development of multi-sequence specific RNAi treatments. And while accounting for all possible combinations across the common mutational positions may be infeasible, perhaps only a threshold is required to supplement the host immune response, providing further detail to the limitations of RNAi. The identification of proportionally dominant genotypes may also aid in such efforts, provided those quasi species remain dominant across more than three generations (S3–S5 Fig).
Altogether this work is focused on mutational patterns in a region of the DWV genome with relatively low nucleotide diversity, we demonstrate that there are specific locations prone to predominantly silent mutations. We also show that the accumulation of mutations may require several successive transmissions to achieve a population of quasi-species capable of evading host immune responses. By using a cloned virus representing a single quasi-species and comparing its mutational progression to the larger population, we gain higher resolution regarding viral dynamics in the host-virus system.
Methods
Honey bees
Three experimental honey bee colonies (here labeled A, B, and C) maintained at the USDA-ARS Bee Research Laboratory in Beltsville, MD were used for source specimens. Colonies were regularly inspected and treated as necessary to remain below a 1% Varroa destructor infection rate. Brood frames were collected from each colony for differential rounds of infection, and a total of 64 pink-eyed pupae were extracted.
Injection groups
All pupae were inoculated using a NanoLuc-tagged DWV clone [18] with an automated 20-gauge needle/syringe into the outer abdomen. Injected pupae were incubated at 35°C and 33% humidity for 72 hours and reared using established methods [16]. Blackened pupae, indicating handling damage or acute infection, were discarded. The injection consisted of a schedule of three successive rounds: first round injections use 106 NanoLuc-tagged DWV clone viral particles as inoculum (F0, 3 pupae). Second-round injections contained DWV isolated from the pooled homogenized solution of round one pupae (F1, 6 pupae per colony). DWV solutions were passed through a 0.2 μm filter to remove cellular material and larger microbes. Third-round injections were prepared in the same way from the second-round homogenate and injected either back into pupae from the same colony or pupae from a different colony (F2, 10 pupae per colony cross) (Fig 1).
Sequencing
Sample definitions were based on the injection generation (F0, F1, F2), source colony and recipient colony. RNA was extracted from infected pupae using the QIAmp Viral RNA Mini kit and converted using SuperScript Reverse Transcriptase oligo(dT). For each library, viral loads were quantified using qPCR on a BioRad CFX-96 thermal cycler as in [33]), using primers designed for both NanoLuc-tagged DWV and wild-type DWV strains [34]. Nanoluc clones were amplified with the forward primer NanoDWVP1.F (CCTGCTGTTCCGAGTAACCA-3’) and the wild-type virus was amplified using DWVP3.F (CCTAATCCAGGACCTGATGGC-3’). Both products were amplified using reverse primer DWVP2.R (TCGCCAATATGCGTGTACCT −3’). Amplicons were quantified by comparing threshold cycles against honey bee actin (GB44311) amplified in parallel with primers Actin.F (TTGTATGCCAACACTGTCCTTT) and Actin.R (TGGCGCGATGATCTTAATTT). Amplicons for a subset of products were purified and sent to the University of Maryland Institute for Genome Sciences for deep sequencing of isolated cDNA using an Illumina MiSeq. The resulting short-read data represents the DWV structural viral protein VP2 that follows the leader protein. Libraries were sequenced for 1 F0 injection, 3 F1 injections, and 6 F2 injections, for each DWV strain. Sequenced reads are available on NCBI under BioProject ID: PRJNA1295475.
Data analysis
Raw sequence read data was aligned to a reference sequence derived from the source viral capsid protein with the bowtie2 aligner [35], using the alignment parameters from Ryabov et. al. 2019 [16]. All additional data analysis was performed using R programming language 4.3.2 Eye Holes [36]. Sequence alignment map files per sample were used to determine the number of reads representing each mismatched base called at each genomic position. The resulting mismatch counts were normalized by rarefaction to the lowest sample read count. To further obtain accurate and comparable mismatch identification across samples, mismatches with less than a liberal estimate of a 1% sequencing error rate were removed from the analysis after normalization. High confidence mismatches were then translated to determine their codon position, amino acid change, and mutation type (missense, silent, or nonsense) using custom R scripts (Supplemental). The ultimate counts of unique mutations and their normalized read representations were compared using Wilcoxon Rank Sum Tests between inoculum groups and their types of mutations, excluding nonsense mutations due to limited representation across samples. Hierarchical clustering of the mismatch counts and positions per sample was performed using the hclust function of the stats package as part of base R and plotted using the ggdendro package [37]. All additional plots were generated with the ggplot2 package [38]. Reference-free calculation of amplicon sequence variants and their proportional representations were performed using the dada2 package [39]. Multisequence alignment was performed with MAFFT v7.526 [40]. The subsequent maximum likelihood trees calculated with RAxML v8 [41], over 1000 bootstrap replicates. The resulting trees were visualized with the ggtree [42] and ggtreeExtra [43] packages.
Supporting information
S1 Fig. Relative transcript abundance (scaled to honey bee transcripts for β-actin) for NanoLuc clone-specific (top) and wild-type primers (bottom) by individual cross of third-round injection cross fostering experiment.
https://doi.org/10.1371/journal.pone.0337191.s001
(PDF)
S2 Fig. Mean number of amplicon sequence variants by strain and injection generation for three levels of proportional representation: > 1%, > 0.1%, and no filter (total).
https://doi.org/10.1371/journal.pone.0337191.s002
(PDF)
S3 Fig. Maximum likelihood tree generated from multisequence alignment of amplicon sequence variants (ASV) with >1% proportional representation.
Legend further identifies ASVs by their strain, injection generation, cross fostering status, and precise proportional representation.
https://doi.org/10.1371/journal.pone.0337191.s003
(PDF)
S4 Fig. Maximum likelihood tree generated from multisequence alignment of amplicon sequence variants (ASV) with >0.1% proportional representation.
Legend further identifies ASVs by their strain, injection generation, cross fostering status, and precise proportional representation.
https://doi.org/10.1371/journal.pone.0337191.s004
(PDF)
S5 Fig. Maximum likelihood tree generated from multisequence alignment of all amplicon sequence variants (ASV).
Legend further identifies ASVs by their strain, injection generation, cross fostering status, and precise proportional representation.
https://doi.org/10.1371/journal.pone.0337191.s005
(PDF)
S1 Table. Nucleotide changes per sample per base.
Data includes the position, normalized count of mutation at each position, original codon and translated amino acid, new codon and translated amino acid, mutation position within codon, mutation type, proportional read representation, and associated metadata.
https://doi.org/10.1371/journal.pone.0337191.s006
(CSV)
S2 Table. Counts and library sizes for sequenced amplicons by strain, injection generation, source colony, end colony, and cross fostering status.
Counts of unique amplicon sequence variants are broken down into three categories of proportional representation: > 1%, > 0.1%, and no filter (total).
https://doi.org/10.1371/journal.pone.0337191.s007
(PDF)
S1 File. Primers – Forward and reverse primers for both the NanoLuc clone and wild-type DWV strains.
https://doi.org/10.1371/journal.pone.0337191.s008
(DOCX)
S2 File. Rcode – Custom code written the R programming language that will translate an open reading frame for an existing sequence, then compare it to a data frame of nucleotide polymorphisms at specific locations, and retranslate the amino acid changes into a new data frame.
https://doi.org/10.1371/journal.pone.0337191.s009
(R)
S3 File. Fasta – Base sequences for both the NanoLuc clone and wild-type DWV strains that were used to compare and derive single nucleotide polymorphisms.
https://doi.org/10.1371/journal.pone.0337191.s010
(FASTA)
Acknowledgments
We thank Karen Shelton and the Charles Herbert Flowers High School Science and Technology Program for support.
References
- 1. Aurell D, Bruckner S, Wilson M, Steinhauer N, Williams GR. A national survey of managed honey bee colony losses in the USA: results from the bee informed partnership for 2020–21 and 2021–22. J Apicultural Res. 2023;63(1):1–14.
- 2.
Steinhauer N, Wilson M, Aurell D, Bruckner S, Williams G. United States honey bee colony losses 2022–2023: preliminary results from the Bee Informed Partnership. 2023.
- 3. Branchiccela B, Castelli L, Corona M, Díaz-Cetti S, Invernizzi C, Martínez de la Escalera G, et al. Impact of nutritional stress on the honeybee colony health. Sci Rep. 2019;9(1):10156. pmid:31300738
- 4. Insolia L, Molinari R, Rogers SR, Williams GR, Chiaromonte F, Calovi M. Honey bee colony loss linked to parasites, pesticides and extreme weather across the United States. Sci Rep. 2022;12(1):20787. pmid:36456591
- 5. Shepherd S, Park Y g, Krupke CH. Effects of common co-occurring pesticides (a neonicotinoid and fungicide) on honey bee colony health in a semi-field study. Heliyon. 2024;10(9).
- 6. Steinhauer N, Kulhanek K, Antúnez K, Human H, Chantawannakul P, Chauzat M-P, et al. Drivers of colony losses. Curr Opin Insect Sci. 2018;26:142–8. pmid:29764654
- 7. Highfield AC, El Nagar A, Mackinder LCM, Noël LM-LJ, Hall MJ, Martin SJ, et al. Deformed wing virus implicated in overwintering honeybee colony losses. Appl Environ Microbiol. 2009;75(22):7212–20. pmid:19783750
- 8. Martin SJ. The role of Varroa and viral pathogens in the collapse of honeybee colonies: a modelling approach. J Appl Ecol. 2001;38:1082–93.
- 9. McMenamin AJ, Genersch E. Honey bee colony losses and associated viruses. Curr Opin Insect Sci. 2015;8:121–9. pmid:32846659
- 10. Dainat B, Evans JD, Chen YP, Gauthier L, Neumann P. Dead or alive: deformed wing virus and Varroa destructor reduce the life span of winter honeybees. Appl Environ Microbiol. 2012;78(4):981–7. pmid:22179240
- 11. Francis RM, Nielsen SL, Kryger P. Varroa-virus interaction in collapsing honey bee colonies. PLoS One. 2013;8(3):e57540. pmid:23526946
- 12. Li N, Li C, Hu T, Li J, Zhou H, Ji J, et al. Nationwide genomic surveillance reveals the prevalence and evolution of honeybee viruses in China. Microbiome. 2023;11(1):6. pmid:36631833
- 13. Grindrod I, Kevill JL, Villalobos EM, Schroeder DC, Martin SJ. Ten years of deformed wing virus (DWV) in hawaiian honey bees (Apis mellifera), the dominant DWV-A variant is potentially being replaced by variants with a DWV-B coding sequence. Viruses. 2021;13(6):969. pmid:34073733
- 14. Wilfert L, Long G, Leggett HC, Schmid-Hempel P, Butlin R, Martin SJM, et al. Deformed wing virus is a recent global epidemic in honeybees driven by Varroa mites. Science. 2016;351(6273):594–7. pmid:26912700
- 15. Doublet V, Oddie MAY, Mondet F, Forsgren E, Dahle B, Furuseth-Hansen E, et al. Shift in virus composition in honeybees (Apis mellifera) following worldwide invasion by the parasitic mite and virus vector Varroa destructor. R Soc Open Sci. 2024;11(1):231529. pmid:38204792
- 16. Ryabov EV, Childers AK, Lopez D, Grubbs K, Posada-Florez F, Weaver D, et al. Dynamic evolution in the key honey bee pathogen deformed wing virus: Novel insights into virulence and competition using reverse genetics. PLoS Biol. 2019;17(10):e3000502. pmid:31600204
- 17. Ryabov EV, Christmon K, Heerman MC, Posada-Florez F, Harrison RL, Chen Y, et al. Development of a honey bee RNA virus vector based on the genome of a deformed wing virus. Viruses. 2020;12(4):374. pmid:32231059
- 18. Evans JD, Banmeke O, Palmer-Young EC, Chen Y, Ryabov EV. Beeporter: Tools for high-throughput analyses of pollinator-virus infections. Mol Ecol Resour. 2022;22(3):978–87. pmid:34612590
- 19. Evans JD, Pettis JS. Colony-level impacts of immune responsiveness in honey bees, Apis mellifera. Evolution. 2005;59(10):2270–4. pmid:16405170
- 20. Evans JD. Transcriptional immune responses by honey bee larvae during invasion by the bacterial pathogen, Paenibacillus larvae. J Invertebr Pathol. 2004;85(2):105–11. pmid:15050840
- 21. Brutscher LM, Daughenbaugh KF, Flenniken ML. Virus and dsRNA-triggered transcriptional responses reveal key components of honey bee antiviral defense. Sci Rep. 2017;7(1):6448. pmid:28743868
- 22. Lu RX, Bhatia S, Simone-Finstrom M, Rueppell O. Quantitative trait loci mapping for survival of virus infection and virus levels in honey bees. Infect Genet Evol. 2023;116:105534. pmid:38036199
- 23. Galbraith DA, Yang X, Niño EL, Yi S, Grozinger C. Parallel epigenomic and transcriptomic responses to viral infection in honey bees (Apis mellifera). PLoS Pathog. 2015;11(3):e1004713. pmid:25811620
- 24. Ray AM, Davis SL, Rasgon JL, Grozinger CM. Simulated vector transmission differentially influences dynamics of two viral variants of deformed wing virus in honey bees (Apis mellifera). J Gen Virol. 2021;102(11):001687. pmid:34816791
- 25. Smeele ZE, Baty JW, Lester PJ. Effects of deformed wing virus-targeting dsRNA on viral loads in bees parasitised and non-parasitised by varroa destructor. Viruses. 2023;15(11):2259. pmid:38005935
- 26. Desai SD, Eu Y-J, Whyard S, Currie RW. Reduction in deformed wing virus infection in larval and adult honey bees (Apis mellifera L.) by double-stranded RNA ingestion. Insect Mol Biol. 2012;21(4):446–55. pmid:22690671
- 27. Durand AM, Bonjour-Dalmon A, Dubois E. Viral Co-Infections and Antiviral Immunity in Honey Bees. Viruses. 2023;15(5):1217. pmid:37243302
- 28. Willemsen A, Zwart MP. Erratum: On the stability of sequences inserted into viral genomes. Virus Evol. 2020;6(1):veaa002. pmid:32015900
- 29. Willemsen A, Zwart MP. On the stability of sequences inserted into viral genomes. Virus Evol. 2019;5(2):vez045. pmid:31741748
- 30. Fuxa JR, Richter AR. Virulence and multigeneration passage of a nuclear polyhedrosis virus selected for an increased rate of vertical transmission. Biological Control. 1992;2(3):171–5.
- 31. Di Prisco G, Cavaliere V, Annoscia D, Varricchio P, Caprio E, Nazzi F, et al. Neonicotinoid clothianidin adversely affects insect immunity and promotes replication of a viral pathogen in honey bees. Proc Natl Acad Sci U S A. 2013;110(46):18466–71. pmid:24145453
- 32. Prisco GD, Zhang X, Pennacchio F, Caprio E, Li J, Evans JD, et al. Dynamics of persistent and acute deformed wing virus infections in honey bees, Apis mellifera. Viruses. 2011;3(12):2425–41. pmid:22355447
- 33. Boncristiani DL, Tauber JP, Palmer-Young EC, Cao L, Collins W, Grubbs K, et al. Impacts of Diverse Natural Products on Honey Bee Viral loads and health. Appl Sci. 2021;11(22):10732.
- 34. Ryabov EV, Posada-Florez F, Rogers C, Lamas ZS, Evans JD, Chen Y, et al. The vectoring competence of the mite Varroa destructor for deformed wing virus of honey bees is dynamic and affects survival of the mite. Front Insect Sci. 2022;2:931352. pmid:38468796
- 35. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9(4):357–9. pmid:22388286
- 36.
Team RC. R: A language and environment for statistical computing. R Foundation for Statistical Computing. 2025.
- 37.
De Vries A, Ripley BD. ggdendro: Create dendrograms and tree diagrams using ‘ggplot2’. R package version 01-20. 2016.
- 38.
Kassambara A, Kosinski M, Biecek P, Fabian S. survminer: Drawing Survival Curves using ‘ggplot2’. R package version 0.4.9. 2021. https://CRAN.R-project.org/package=survminer
- 39. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13(7):581–3. pmid:27214047
- 40. Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol. 2013;30(4):772–80. pmid:23329690
- 41. Kozlov AM, Darriba D, Flouri T, Morel B, Stamatakis A. RAxML-NG: a fast, scalable and user-friendly tool for maximum likelihood phylogenetic inference. Bioinformatics. 2019;35(21):4453–5. pmid:31070718
- 42. Yu G, Smith DK, Zhu H, Guan Y, Lam TT. ggtree: an r package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods Ecol Evol. 2016;8(1):28–36.
- 43. Xu S, Dai Z, Guo P, Fu X, Liu S, Zhou L, et al. ggtreeExtra: compact visualization of richly annotated phylogenetic data. Mol Biol Evol. 2021;38(9):4039–42. pmid:34097064