The Merle Gene in Dogs: Genetics, Health Risk and Welfare

1. What Merle Is, Genetically

1.1 The Mutation

Merle results from the insertion of a short interspersed element — a mobile piece of DNA roughly 253 base pairs long — at the boundary between intron 10 and exon 11 of the PMEL gene, formerly designated SILV or PMEL17, on canine chromosome 10 (Clark, Wahl, Rees & Murphy, 2006).

This is worth stating precisely, because merle is frequently described as a point mutation. It is a retrotransposon insertion, and that structural difference is what produces the variable behavior described in the next section.

1.2 What It Does to Pigment

The insertion disrupts melanosome function and thereby the distribution of eumelanin — the black and brown pigments — in coat, skin and eye. Areas dominated by phaeomelanin, the red and yellow pigments, show little or no visible merle patterning.

The practical consequence is immediate: merle is conspicuous on a black dog and can be nearly invisible on a red one. Coat colour interacts with the genotype to determine what an observer sees, which is the first of several reasons appearance is a poor guide.

1.3 Incomplete Dominance

Merle is often described as simply dominant. It is more accurately incompletely dominant. One copy produces the pattern; two copies do not produce more pattern but a qualitatively different and considerably higher-risk phenotype.

That distinction is not academic. It is the entire reason merle-to-merle pairings are treated as a welfare issue while merle-to-non-merle pairings generally are not (a case where genetics rather than behavior determines outcomes).

2. Merle Is a Spectrum, Not a Switch

2.1 Length Determines Effect

The inserted sequence carries a poly-A tail whose length varies between individuals, and that length co-determines how strongly merle expresses. Across merle breeds, a length spectrum from roughly 200 to 280 base pairs has been described and divided into several categories, with the authors explicitly describing a graded risk of health impairment (Langevin, Synkova, Jancuskova & Pekova, 2018).

2.2 The Variant Categories

Working from the shortest upward: the wild-type allele carries no insertion and produces no merle. Cryptic merle alleles, in the low-200s, produce no visible pattern at all. Cryptic-plus alleles, running into the mid-240s, likewise produce a solid-looking coat. Atypical merle, in the high 240s to low 250s, produces no pattern but sometimes a lightened or brownish tone. Atypical-plus, into the low 260s, can produce weak patterning and a muted, ill-defined colour. Classic merle sits in the mid-to-high 260s and produces the recognisable pattern. Harlequin merle, from around 269 upward, produces highly variable patterns including extensive dilution toward white.

Two cautions belong with that list. The boundaries are practical classifications used by testing laboratories rather than sharp biological thresholds, and the transitions between categories are gradual. The allele set, the naming and the overall length range come from the published work; the exact cut-off values are laboratory conventions.

2.3 The Consequence

A dog can carry merle without looking like a merle. Cryptic and atypical variants produce little or no visible pattern, they are inherited normally, and in the following generation they can meet a long allele. The mostly white deaf puppy from two apparently non-merle parents is not a mystery — it is this mechanism working as expected.

3. Mosaicism: Why Appearance Proves Nothing

3.1 The Finding

The poly-A tail is a highly repetitive sequence and therefore prone to replication error during cell division. The result is that a single dog can carry different merle alleles in different body cells. In the published sample, this genetic mosaicism was found in 30 of 181 dogs — 16.6 percent (Langevin et al., 2018).

3.2 Why It Matters for Breeding

In a mosaic animal, different cell lines carry different merle variants, and which variant is transmitted depends on which cell line gives rise to the germ cells. In males in particular, examining sperm can become relevant, because germline variants may differ from a blood test result — which explains litters that otherwise appear inexplicable.

The authors' conclusion was procedural: mosaic findings must be reported in full to the breeder rather than simplified away.

3.3 What Follows for Owners and Buyers

A breeding partner that "does not look merle" is not evidence of anything. Genetic testing is currently the most reliable way to identify merle variants, though results should still be interpreted in the context of the breeding line (a reminder that visible traits are unreliable proxies generally).

4. Deafness: The Measured Rates

4.1 The Study

The most robust data come from an objective hearing assessment of 153 merle dogs across breeds, with genotype determined in parallel (Strain, Clark, Wahl, Turner & Murphy, 2009).

Among single merle dogs, 2.7 percent were unilaterally deaf and 0.9 percent bilaterally deaf — around 3.5 percent affected in total. Among double merles, 10 percent were unilaterally deaf and 15 percent bilaterally deaf, for roughly 25 percent affected. The association between hearing status and genotype was statistically significant. There was no significant association with eye colour or with sex.

4.2 What This Means for Single Merles

The common claim that single merle carries no health consequence is a simplification. The risk is small but not zero. The overwhelming majority of single merle dogs hear and see normally, which makes a hearing test in puppies worthwhile precisely because unilateral deafness is easily missed in daily life for years.

4.3 Calibration Against Other Breeds

Deafness prevalence in merle dogs exceeded that in some breeds homozygous for the piebald gene, such as the English Cocker Spaniel, but was comparable to or lower than that in Dalmatians and white Bull Terriers (Strain et al., 2009). Merle is one instance of a broader pattern in which reduced pigmentation in sensory tissue carries auditory risk — not a uniquely dangerous gene (with breed-level generalisations warranting caution throughout).

5. Ocular Effects

5.1 The Clinical Picture

A classical case series described 45 Australian Shepherds with microphthalmia and multiple colobomas. Among affected dogs, microphthalmia, microcornea, and heterochromia with dyscoria and corectopia were each present in all cases; cataracts were found in 62 percent of affected eyes, equatorial staphylomas in 54 percent, and retinal detachments in 53 percent. Affected dogs were 30 to 90 percent white with the remainder blue merle, and cardiac anomalies were noted in six dogs (Gelatt & McGill, 1973).

Read this correctly: these are proportions within already-affected dogs, not prevalence figures for merle dogs generally. There is no ocular equivalent of the Strain hearing dataset.

5.2 Inheritance

Follow-up work characterised microphthalmia with coloboma in the merle Australian Shepherd as an incompletely penetrant recessive trait (Gelatt, Powell & Huston, 1981). Incomplete penetrance means that carrying the relevant genotype does not guarantee the phenotype, which is part of why outcomes in double merle litters vary so widely.

5.3 Pigment and Severity

The most severe ocular findings occur in homozygous merles with extensive white on the head. Dogs with substantial unpigmented skin also lack the protection pigment normally provides, which makes sun exposure a practical management issue rather than a cosmetic one. A dog that avoids bright conditions may be responding to discomfort rather than to the situation (a reading error that recurs across behavior problems).

6. Breeding, Law and Testing

6.1 The German Legal Frame

Section 11b of the German Animal Welfare Act prohibits breeding vertebrates where it is to be expected that offspring will hereditarily lack body parts or organs, or have them unsuited or altered, in ways that cause pain, suffering or damage.

Applied to merle, the deliberate pairing of two merle carriers is the constellation in which impaired offspring are to be expected. This is a description of the statutory test, not legal advice; specific breeding questions belong with the responsible veterinary authority.

6.2 What Reduces Risk

Merle to non-merle substantially reduces the risk of a double merle outcome — but only if the genetic variants of both parents are actually known. Given cryptic and atypical alleles, "looks non-merle" does not establish "is non-merle," which is why testing rather than inspection is the operative safeguard.

Professional bodies additionally recommend testing dogs with a merle parent or merle siblings before breeding regardless of breed, testing puppies from merle pairings at the M locus before rehoming, and informing prospective owners of intact merle dogs about the risks in writing.

6.3 Recent Introductions Are the Higher-Risk Case

Merle is long-established in a number of breeds, including the Australian Shepherd, Border Collie, Collie in both coat types, Shetland Sheepdog, Cardigan Welsh Corgi, Dachshund, Great Dane in connection with harlequin, and the Catahoula Leopard Dog. In others — French Bulldog, Chihuahua, Pomeranian, and various bully types — it appears through recent outcrossing driven by market demand.

The recent introductions are treated as the more critical situation, because the gene's effects differ between breeds and newly introduced lines have no accumulated experience to draw on. No complete list of merle-carrying breeds exists, and through mixed breeding merle can appear anywhere.

7. Living and Training with Sensory Impairment

7.1 Most Merle Dogs Need Nothing Special

A single merle with intact hearing and normal vision is an ordinary dog. What is worth doing anyway: a hearing test in puppies from merle pairings, annual eye and ear checks, and sun protection where skin is unpigmented. Early detection matters because a puppy with undiagnosed unilateral deafness moves through its socialisation window with incomplete information (during the period when that matters most).

One point is easy to overlook. A dog with reduced perception may register or report discomfort differently, which makes behavioral change worth taking seriously (since pain often presents behaviorally).

7.2 Learning Is Unchanged; the Channel Changes

A dog that cannot see or hear well learns exactly as any other dog does. Learning principles are unaffected — what changes is the channel through which information arrives (the underlying mechanisms being identical).

The governing principle is to announce rather than surprise. Reduced warning time is what produces startle responses, withdrawal or defensive behavior in these dogs — not temperament (which is a state effect rather than a trait one). Repeated unannounced contact is also the reliable way to build a conditioned response to being touched (by exactly the mechanism fear learning describes).

7.3 Working Channels

Scent is the strongest remaining channel and the most rewarding to work with (given how central olfaction is). Consistent scent markers at fixed points — sleeping place, water bowl, the door to outside — build a navigable map.

Touch functions as a full signal system when executed consistently: a flat hand on the chest for stop, a touch at the croup for move on, a tap at the shoulder for a direction change. What matters is that each signal always arrives at the same place, with the same intensity, meaning the same thing — and that it is introduced while the dog is calm enough to process it (which the arousal literature makes a precondition). Build them as any other behavior — shape first, consolidate, then add the tactile cue — rather than the reverse.

Abrupt leash pressure is explicitly not a tactile cue. A signal has to be discriminable, not unpleasant, and a dog with less warning than usual is the last one that should receive an input it cannot see coming (which is where aversive input does its damage).

7.4 Environment and Social Contact

Keep furniture and routes stable, maintain predictable feeding and walking times (since unpredictability is itself a stressor), and treat off-lead freedom as a genuine risk — a long line on a well-fitted Y-harness preserves movement without losing control (predictability being a documented buffer against anxiety).

A dog that cannot fully read other dogs' signals benefits from calm, confident partners and suffers under pushy ones (since signal reading is itself learned). Quality clearly beats quantity here, and a stable relationship with the handler carries more weight than usual (as the attachment literature indicates).

8. Summary at a Glance

Merle is a retrotransposon insertion, not a point mutation — A SINE of roughly 253 base pairs at the intron 10 / exon 11 boundary of PMEL on chromosome 10 (Clark et al., 2006).

Incompletely dominant — One copy gives the pattern; two give a qualitatively different and higher-risk phenotype, which is why merle-to-merle is the problem case.

A spectrum, not a switch — Poly-A tail length from roughly 200 to 280 base pairs sorts into cryptic, atypical, classic and harlequin variants with graded risk (Langevin et al., 2018).

Carriers can look unremarkable — Cryptic and atypical alleles produce little or no visible pattern and are inherited normally.

Mosaicism was found in 16.6 percent — 30 of 181 dogs carried different merle alleles in different cells, so a blood test may differ from what is transmitted (Langevin et al., 2018).

Deafness: 3.5 percent versus 25 percent — Single merles 2.7 percent unilateral and 0.9 percent bilateral; double merles 10 percent unilateral and 15 percent bilateral, across 153 dogs (Strain et al., 2009).

No eye-colour or sex association — Neither predicted hearing status (Strain et al., 2009).

Ocular findings are severe but not quantified as prevalence — The classical series describes what affected dogs have (Gelatt & McGill, 1973), not how many merle dogs are affected.

9. Research Gaps and Critical Appraisal

There is no ocular equivalent of the deafness dataset. The Strain study gives genotype-linked prevalence for hearing. Nothing comparable exists for eyes: the ocular literature is built on case series of affected animals, which characterises severity well and prevalence not at all.

The allele cut-offs are laboratory conventions. The variant set, the naming and the overall length range are published. The precise base-pair boundaries between categories are practical classifications used by testing laboratories, and the biological transitions are gradual rather than stepped.

Breed-specific risk is largely unquantified. The deafness sample spanned multiple breeds, and effects are known to differ between breeds — but breed-level figures are unavailable for most, which is precisely the situation in recently outcrossed lines where risk assessment matters most.

The core ocular work is over four decades old. Gelatt and McGill's series dates from 1973 and the inheritance study from 1981, both predating molecular characterisation of the locus. They have not been superseded by comparable modern work.

Mosaicism has been described but not systematically quantified across breeds. A single study established the phenomenon and its frequency in one sample. How it varies between populations, and how often germline and somatic results diverge in practice, remains open.

Welfare outcomes for affected dogs are not systematically tracked. How double merle dogs with sensory impairment actually fare over a lifetime — behaviorally, medically, in terms of rehoming success — is not documented in any structured way (and behavioral outcomes are hard to read directly).

10. Conclusion

Merle is neither harmless nor catastrophic; it is a breeding decision with quantified consequences, and the numbers are specific enough to argue from. A single merle dog carries a small but non-zero risk of deafness at roughly 3.5 percent, which makes the common reassurance that single merle is entirely safe a simplification rather than a fact. A double merle carries around 25 percent, with 15 percent bilaterally deaf, alongside ocular malformations that can be severe. What turns this from a genetics curiosity into a practical welfare problem is that appearance does not identify carriers: cryptic and atypical variants produce dogs that look entirely ordinary and transmit merle normally, and genetic mosaicism means even a test result reflects the cells sampled rather than necessarily the cells that will produce offspring. The operational conclusions follow directly. Test rather than inspect, ask for both parents' genotypes before buying, and treat a refusal to produce them as the answer it is. For the dogs already living with impairment, the picture is more ordinary than it sounds — learning principles are unchanged, only the channel differs, and a dog navigating by scent and touch on predictable routes is not a diminished dog but one working with a different set of inputs (much as counterconditioning works through whichever channel is available).

Key Insights (Takeaways)

• A dog can carry merle without looking like one. Cryptic and atypical variants of the insertion produce little or no visible pattern while being inherited normally (Langevin et al., 2018), which is how two apparently non-merle parents produce an affected puppy. Inspection is not a substitute for a genetic test, and this single fact drives most of the practical guidance in this area.

• The deafness difference between genotypes is large and measured. Across 153 dogs, 3.5 percent of single merles were deaf in at least one ear, against 25 percent of double merles, with 15 percent of double merles deaf bilaterally (Strain et al., 2009). Eye colour and sex predicted nothing. The claim that single merle carries no risk is a simplification; the risk is small, not absent.

• Merle is incompletely dominant, and that is the whole argument. Two copies do not produce more pattern but a qualitatively different phenotype with substantially higher risk — which is why merle-to-merle pairing, rather than merle itself, is the welfare issue.

• Genetic mosaicism was found in 16.6 percent of a sampled population. Thirty of 181 dogs carried different merle alleles in different cells (Langevin et al., 2018), meaning a blood test may not reflect what is transmitted through the germline — and that mosaic findings must be reported to breeders in full rather than simplified.

• The ocular evidence describes severity, not prevalence. The classical series documents what affected dogs have — microphthalmia and microcornea in all cases, cataracts in 62 percent of affected eyes, retinal detachment in 53 percent (Gelatt & McGill, 1973) — but no study gives genotype-linked ocular prevalence the way the hearing data does. Quoting these figures as merle risk rates misreads them.

References

Clark, L. A., Wahl, J. M., Rees, C. A., & Murphy, K. E. (2006). Retrotransposon insertion in SILV is responsible for merle patterning of the domestic dog. Proceedings of the National Academy of Sciences, 103(5), 1376–1381. https://doi.org/10.1073/pnas.0506940103

Gelatt, K. N., & McGill, L. D. (1973). Clinical characteristics of microphthalmia with colobomas of the Australian Shepherd Dog. Journal of the American Veterinary Medical Association, 162(5), 393–396. https://pubmed.ncbi.nlm.nih.gov/4691375/

Gelatt, K. N., Powell, N. G., & Huston, K. (1981). Inheritance of microphthalmia with coloboma in the Australian Shepherd Dog. American Journal of Veterinary Research, 42(10), 1686–1690. https://pubmed.ncbi.nlm.nih.gov/7325429/

Langevin, M., Synkova, H., Jancuskova, T., & Pekova, S. (2018). Merle phenotypes in dogs – SILV SINE insertions from Mc to Mh. PLoS ONE, 13(9), e0198536. https://doi.org/10.1371/journal.pone.0198536

Strain, G. M., Clark, L. A., Wahl, J. M., Turner, A. E., & Murphy, K. E. (2009). Prevalence of deafness in dogs heterozygous or homozygous for the merle allele. Journal of Veterinary Internal Medicine, 23(2), 282–286. https://doi.org/10.1111/j.1939-1676.2008.0257.x