A four-genome view — host, mite, virus, and population — of what genetics tells us about bee health.
🔍 Search the corpus · 🧬 Research Frontier · 🦠 Virus hubs · 🕷️ Varroa · 🏠 Home
A structured deep dive across four genetic angles — host resistance & breeding, the Varroa mite's own genome, viral genomics & DWV evolution, and population genetics, lineages & introgression — built from the archive's evidence packs, the 2023–2026 discoveries scan, and full-text reading of key primary papers. Citations link to each paper's DOI (or PubMed Central). This is a curated synthesis, not an exhaustive systematic review; every single-study claim is flagged and deserves independent confirmation.
Bee health is a four-genome problem: the host bee, the mite, the virus, and the population/lineage context that frames all three. The deep dives below treat each in turn, but the most important insights live where they intersect.
The big picture per angle:
Three threads that tie it together:
The "two-pathway" DWV puzzle (viral genomics × population genetics). Europe/USA went DWV-B-dominant; Latin America/China/Japan stayed DWV-A-dominant. The sharp test is Cuba — isolated 60+ years, untreated, Varroa-resistant — which should look like its DWV-A neighbours but instead mirrors the USA/Europe (DWV-B dominant + recombinant). Meanwhile Africanized Yucatán bees retained DWV-A. Host lineage, founder effects, superinfection exclusion, and Varroa-independent drift are all on the table; nobody knows why. (See §3 and §4; Cuba = Hesketh-Best 2026, n=5, treat as a lead.)
Tolerance vs resistance (host breeding × population genetics × viral evolution). "Survivor" stocks may not be genetically resistant at all — Fernando de Noronha survives 30+ years simply because its DWV quasispecies never evolved a virulent variant (Brettell 2017), and a 2025 critique warns survival-based "Darwinian" beekeeping selects for tolerance that sustains pathogen reservoirs and may select for greater virulence in wild bees (Sokolov 2025). Reducing mite numbers and reducing virus harm are genetically distinct goals.
Varroa as an evolutionary force, not just a parasite (mite × host population genetics). On the Azores, mite-infested islands carry far higher commercial-lineage introgression than mite-free islands with the same import history — Varroa independently raised admixture by +0.118 Q (Henriques 2025). The mite reshapes the host gene pool by killing colonies that then get restocked with commercial queens.
What's genuinely new (2023–2026): the SOV trait (breeding directly for heritable virus resistance, decoupled from the mite); host antiviral targets (arginine kinase, snapin, m6A/METTL3, AF9); RNAi maturing into a registered-stage acaricide (vadescana, with strong bee-safety data); the Laelapidae reclassification; the BABA/BABB "fourth genome block" recombinant; and the recasting of Varroa as a bacterial vector (Morganella morganii).
Varroa destructor switched from Apis cerana to A. mellifera in the mid-20th century and, by vectoring and amplifying DWV, became the dominant cause of colony loss; untreated European colonies typically collapse within 2–3 years (Behrens 2011). Acaricide dependence has both masked and suppressed natural selection for resistance. Because the colony is the unit of selection but workers express the behaviours, resistance genetics is unusually complex: largely behavioural social immunity, polygenic, and strongly population-specific.
Four behavioural/physiological traits underpin most documented resistance, and they overlap rather than being independent:
Genetic architecture: polygenic with small-effect QTL and strong epistasis. The clearest mapping (Behrens 2011, Gotland-tolerant × susceptible drones): three QTL suppressing mite reproduction on chromosomes 4, 7, 9, individually weak (chr 7 = 8.7% of variance, chr 4 = 5.3%, chr 9 = 3.7%) but jointly significant through epistasis — which the authors flag as complicating marker-assisted breeding. An ecdysone-induced gene recurs at resistance loci across independent stocks (Conlon 2019) — biologically attractive because mites cannot synthesise ecdysone but acquire it from the host, which times their reproduction. Heritabilities are low-to-moderate and trait-dependent, so field testing remains essential.
The most-used marker is the chromosome-9 SNP 9-9224292 (A/G); homozygous-G A. m. ligustica colonies had 28.5% lower mite levels — real but insufficient to replace treatment (Sainsbury 2022, Sainsbury 2022). The population-specificity problem is the field's biggest obstacle: the opposite allele associates with VSH in Russian bees (Kirrane 2014, Kirrane 2014), and the eight mite-non-reproduction variants from a single Dutch colony have allele frequencies that track lineage (M vs C) (Lefebre 2024) — markers don't transfer across races without re-validation. A new mean-reproduction-rate metric (mVR) reaches DMR-equivalent precision with 10 infested cells vs 40 (Lefebre 2024), cutting breeding labour. (Confidence: high that MAS is real but population-specific.)
The weight of evidence places resistance in behaviour mediated by olfaction/chemosensation (tolerant bees differ mainly in neuronal/olfaction genes — Navajas 2008; VSH-discriminating transcription localises to the antennae — Mondet 2015, Mondet 2015), secondarily in physiology/immunity (ecdysone pathway, haemocytes/AMPs), and least in classical antibacterial immunity. The covert-DWV transcriptome (Kara 2026) cautions that low-load infection upregulates venom/cuticular genes, not classical immunity — so "immune gene expression" is a poor universal readout.
Varroa host-shifted from A. cerana; on naïve A. mellifera, untreated colonies collapse within 2–3 years (Garbian 2012). The global pandemic descends from very few mitochondrial haplotypes (chiefly "Korea", secondarily "Japan"), so the worldwide mite population carries strikingly low nuclear diversity — a recurring theme for resistance evolution. The major recent revision: a combined molecular phylogeny (13 mt protein-coding genes across 24 taxa + 4 nuclear loci across 113 taxa) nests Varroa within Laelapidae, demoting Varroidae to subfamily Varroinae stat. nov. (Oh 2024). Bee-associated lifestyles evolved independently ≥3× within Laelapidae. (Confidence: high for the topology; the nomenclatural change is still propagating.)
The foundress times oogenesis to host cues and cannot synthesise ecdysone (acquires it from the pupal meal) — so mite reproduction is a calcium- and hormone-gated developmental program. RNAi knockdown of Varroa calmodulin abolishes reproduction via disrupted calcium signalling and cadherin/embryo-development genes; treated cells contain only misshapen, disintegrating eggs by day 5, with foundress survival unaffected (Smeele 2026). Salivary secretions prevent wound healing (the conduit for pathogen injection) and include immunomodulatory effectors (cystatin among candidates), though specific salivary-effector genetics remain underexplored.
South Korean Varroa metatranscriptomics found 16 viruses incl. two newly reported, with DWV-B/VDV-1 actually replicating in mite gut/salivary cells (Kim 2026). The mite also vectors bacteria — a Morganella morganii strain killed >30% of bees at ~215 CFU, transmitting bee→mite 92.1% and mite→bee 68.5% but not bee-to-bee (Chen 2025). The mite microbiome is more disruptable than the bee's, making the mite holobiont itself a candidate target. (Confidence: medium — pre-validation.)
Low post-host-shift diversity is double-edged (fast resistance fixation, but ideal conserved-gene RNAi target); is amitraz resistance one locus or many; resistance management is running out of modes of action; mite-resistance and host virus-tolerance are genetically distinct axes (Thaduri 2019); salivary-effector genetics are the least-characterised next RNAi target set.
DWV is a +ssRNA Iflavirus; reference genome 10,140 nt, single ORF → ~328-kDa polyprotein (capsids VP1/VP2/VP3 at the N-terminus; helicase, 3C protease, RdRp at the C-terminus), with a long 5′ leader carrying an IRES (Lanzi/Genersch, Lanzi 2006). DWV is a quasispecies resolving into master variants: common DWV-A, DWV-B (formerly VDV-1), rare DWV-C, and apparently extinct DWV-D (Hesketh-Best 2026). A and B share only ~84% genome-wide identity (within-variant ~96–99%); now treated as strains of one species (Iflavirus aladeformis), a reclassification justified by pervasive A×B recombination (Dalmon 2017). DWV's transformation into the world's most prevalent insect virus was driven by Varroa injecting it into the pupal haemocoel, bypassing the oral/vertical routes that kept it benign.
Recombination is arguably the dominant force shaping DWV genomes. Dalmon (Dalmon 2017) found nine breakpoints, non-random, with two hotspots (5′UTR/leader-protein region; conserved helicase region) at the boundaries of three functional modules (5′UTR; structural; non-structural) — matching the modular architecture of other Picornavirales. A recurring pattern is a DWV-A non-structural (replication) module + DWV-B capsid module: such recombinants accumulate higher than either parent and correlate between pupae and their mites (better adapted to the bee↔mite cycle). US surveys now find recombinants dominate. The newest extension: a BABA/BABB recombinant ("fourth genome block") rising in the British Isles/France, and "independent of varroa" presence (Dobelmann 2025) — complicating the mite-centric model. (Confidence: moderate — single recent study.)
The keystone experiment (Ryabov, Ryabov 2014): Varroa-free pupae carry low/diverse DWV; Varroa-associated pupae either low/diverse or high-titre near-clonal virulent (amplified 1,000–10,000×). Mites carry a diverse population, yet transmission from the mite — or direct injection of mixed populations — selects the single virulent variant. The route (haemocoel injection), not the mite per se, advantages it. This is the molecular basis of the Hawaii natural experiment (Varroa arrival collapsed DWV diversity + millionfold titre rise). A two-phase model follows: Varroa-induced sweep/bottleneck, then diversification under negative frequency-dependent selection (rare genotypes evade RNAi).
Consensus held DWV-B is more virulent and rising, supported by (a) DWV-B (not A) infecting mite gut/salivary tissue (Gisder 2021) and (b) DWV-B eliciting a far smaller host transcriptional response (189 vs 2,906 DEGs) despite higher titres — immune evasion (Ray 2025). But matched clone studies contradict the ranking: DWV-B gave 20% pupal mortality vs 80% for DWV-A; at low dose 90% of DWV-B bees were asymptomatic vs 100% DWV-A mortality, converging only at high dose (Barth 2024). This dissociates virulence (per-pupa lethality — DWV-A wins) from epidemiological fitness/transmissibility (DWV-B rising). Reinforced by Yucatán (DWV-B never displaced DWV-A; Fleites-Ayil 2025) and a dysbiosis study where DWV-A dominated (Becchimanzi 2026). (Confidence: the field is genuinely contradictory; clone studies should temper any blanket "DWV-B is more virulent.")
Europe/USA/UK went DWV-B-dominant + recombinant (UK 100%, Europe 82%, US 56%); Latin America/China/Japan stayed DWV-A-dominant (Brazil 11%, Mexico 0.7–2%, China 0.8% DWV-B) (Hesketh-Best 2026). The sharp test: isolated, untreated, Varroa-resistant Cuba (60+ yr import ban) — expected to look DWV-A like its neighbours — instead mirrors the USA/Europe (DWV-B dominant 75%, recombinant present), with BQCV absent and LSV newly detected (Hesketh-Best 2026; n=5, RNA degraded by power cuts — a provocative lead, not settled). Candidate explanations: superinfection exclusion / "recombination meltdown" locking A-dominant regions (untested); transmission bottlenecks + B-biased replication (drone serial-passage, Woodford 2023); Varroa-decoupled drift (the BABA/BABB shift). The authors' own verdict: "how and why two distinct evolutionary DWV pathways have arisen remain a mystery."
Four codons under positive selection (dN/dS>1): three in the leader protein (codons 21, 57, 107) and one at the RdRp 3′ end (codon 2838) (Dalmon 2017) — the leader protein may shut off host cap-dependent translation to favour IRES-driven viral translation. Host-factor work adds candidate antiviral targets: DWV co-opts arginine kinase (Becchimanzi 2025) and the capsid-binding snapin (Sun 2023).
LSV — highly variable, stable (non-emerging) association; Varroa elevates and may be driving its adaptive evolution (Azores); new variants in declining US colonies (Jones 2026) and first detection in Cuba. IAPV/ABPV/KBV — recombination + quasispecies; RNAi genotype-specific. BQCV/SBV — switching oral→injected entry sharply raises virulence + 17–35× more DEGs (Al 2023), a transcriptomic parallel to DWV; BQCV strikingly absent in isolated Cuba; SBV has its own host-resistance genetics (chr-15 SNPs in A. cerana, Hassanyar 2023). EVEs/NIRVS — heritable integrated virus sequences in Bombus with possible piRNA-based antiviral priming (de 2026; weakly supported, Bombus not Apis).
Why two DWV pathways (SIE untested); does recombination raise or lower virulence (contested); is the A→B shift Varroa-driven or partly mite-independent; reconciling virulence vs fitness; is DWV's architecture three blocks or four.
A. mellifera splits into four deep lineages — A (African), M (W/N European incl. mellifera), C (E European/Mediterranean incl. carnica, ligustica), O (Near/Middle Eastern) — robustly recoverable from genome-wide SNPs. The most rigorous recent statement on defining subspecies comes from sister-species A. cerana (Qiu 2023): whole-genome SNPs from 362 workers resolved 8 subspecies and showed morphology (esp. body size) tracks climate, not history — only environment-independent characters (wing venation) recovered genomic boundaries. Proposed criteria: evolutionary independence, trait distinction, geographic isolation — a prerequisite for "strategized conservation." African selection scans (3.6M SNPs, 11 Kenyan workers) confirm lineages are adaptively differentiated with positive-selection signatures (Fuller 2015).
The unifying lesson: resistance arises by standing variation (Africa), bottleneck selection (Gotland), founder/quasispecies contingency (Noronha), or novel single mechanisms (California) — not interchangeable, and only the panmictic route preserves diversity.
The striking recent finding: Varroa drives introgression. On the Azores, mite-free islands retained negligible commercial C-lineage introgression (Q 0.004–0.091) despite the same historical imports, while mite-infested islands carried high introgression (Q 0.156–0.261); Bayesian modelling attributed an independent +0.118 Q increase to Varroa presence (Henriques 2025) — the mechanism being colony loss → restocking with commercial queens. Varroa as an evolutionary force on the host gene pool, not just a vector. (Confidence: medium — regional/observational but well-controlled by the island contrast.) Africanization intersects with viral evolution: AmFV clade 3 dominates African-ancestry bees (Cornman 2023), and Yucatán Africanized bees show DWV-B failing to replace DWV-A (Fleites-Ayil 2025) — host lineage shapes which viral variants prevail.
Cuba: ~220,000 colonies, 60+ year import ban — the largest isolated Varroa-resistant European-derived population (Hesketh-Best 2026). Despite Varroa (1996) + DWV reaching 91% of colonies (2018), it persists untreated. The virome surprise (see §3): DWV evolution mirrors the USA/Europe (DWV-B dominant + recombinant), unlike its Latin American neighbours; BQCV absent; LSV first-detected. The central caveat: only 5 usable samples (power-cut RNA degradation) → 8 contigs. (Confidence: low-to-moderate; hypothesis-generating.)
A dedicated honey-bee immune-gene SNP assay (123 whole genomes, 7 subspecies, 3 lineages → 91 SNPs across 89 innate-immune genes) clustered samples by lineage/subspecies from immune variation alone (Henriques 2021) — ancestry is legible even in immune loci, and immune diversity is partitioned among lineages; monitoring it "can inform on the adaptive potential to resist the invaders." This creates tension with breeding: MAS markers are population-specific (Lefebre 2024), and intensive selection on narrow stocks (or survival-based "Darwinian" beekeeping) risks eroding local-adaptation/immune diversity and selecting tolerance over resistance (Sokolov 2025). The counter-pressure: isolated local gene pools are worth protecting yet chronically eroded by commercial introgression that Varroa accelerates (Henriques 2025).
Is Noronha-style tolerance host adaptation or viral contingency; does the Varroa-drives-introgression effect generalise beyond islands; why do isolated populations (Cuba) recapitulate the global DWV-B trajectory while neighbours retain DWV-A; how to reconcile MAS breeding with diversity conservation (esp. since high diversity did not protect wild bumble bees from disease — Dobelmann 2025); and standardised environment-independent subspecies criteria for A. mellifera are still not uniformly adopted.
Curated synthesis compiled 2026-06-23 from the archive's full-text corpus. Citations resolve to DOI or PubMed Central. Companion page: Research Frontier: Recent Developments.