Most of what the public reads about inbreeding comes down to one rule: low COI is responsible, high COI is not. Breeders don't read the number that way. To a breeder, COI describes what a pairing does. It doesn't grade it.
Every recognized breed exists because foundation breeders bred related dogs to fix type. The German Shepherd was built by linebreeding on one founding sire, Horand von Grafrath. The AKC still calls linebreeding “essential to setting type and solidifying positive characteristics,” with outcrossing as a periodic complement. For the working vocabulary, see the breeding strategies glossary.
The short version
- What it measures: The chance a puppy gets the same copy of a gene from both parents because both inherited it from one ancestor.
- How it's worked out: Each path linking sire and dam through a shared ancestor adds ½ for every dog on it. Half-siblings give 12.5%, full siblings 25%.
- The number names a strategy: Five bands run from outcross (0–3%) to close inbreeding (25%+). The band describes the pairing; it doesn't grade it.
- DNA usually reads higher: A pedigree treats its earliest recorded dogs as unrelated. Embark reports a genomic COI; Wisdom Panel reports a different diversity score.
What a coefficient of inbreeding measures
The coefficient of inbreeding (COI) is the chance that a puppy inherits the same copy of a gene from both parents, because both parents got it from the same ancestor. It is written as a percentage.
Every dog gets half its DNA from its sire and half from its dam. If the sire and dam are related, some of that DNA is the same DNA, handed down both sides of the pedigree. COI puts a number on the overlap.
A COI of 0% means the parents share no ancestors in the pedigree being read. A COI of 25% is what a parent-to-offspring or full-sibling mating gives. Most purebred litters fall between the two.
How COI is worked out from a pedigree
The method is Wright's coefficient of inbreeding, published by the geneticist Sewall Wright in 1922. It finds every path through the pedigree that links the sire and the dam through a shared ancestor.
Each path adds ½ for every dog on it: count the sire, the dam, the shared ancestor and every dog between them, then multiply ½ by itself that many times. Written as a formula, that is 0.5 raised to the power of (n1 + n2 + 1), where n1 and n2 are the generations from the sire and from the dam up to the shared ancestor.
Worked example: half-siblings. The sire and the dam are both by the same dog. One path links them: sire, shared sire, dam. Three dogs, so ½ × ½ × ½ = 12.5%.
Full siblings share both parents, so there are two paths of three dogs each: 12.5% + 12.5% = 25%.
Two rules keep the count honest. A path may not pass through the same dog twice. And if the shared ancestor is itself inbred, its path is multiplied by (1 + that ancestor's own COI).
Our free COI calculator does the count for a 3-generation pedigree. It is a quick estimate: it treats each shared ancestor as not inbred, and when the same dogs repeat at several depths on both sides it can count one line of descent more than once and read high. For a pedigree like that, check the figure against a pedigree database or software that traces each path separately.
COI by relationship
These are the figures for a single relationship. They assume the rest of the pedigree is unrelated and the shared ancestors are not themselves inbred. More shared ancestors further back push the real number higher.
| The mating | Litter's COI | Usually called |
|---|---|---|
| Unrelated | 0% | Outcross |
| Same dog 3 generations back on one side, 4 on the other (a 3-4) | 1.6% | Outcross |
| Same great-grandparent on both sides (a 3-3) | 3.1% | Light linebreeding |
| First cousins | 6.25% | Light linebreeding |
| Half-siblings | 12.5% | Moderate linebreeding |
| Full uncle–niece or aunt–nephew | 12.5% | Moderate linebreeding |
| Grandparent and grandchild | 12.5% | Moderate linebreeding |
| Full siblings | 25% | Close inbreeding |
| Parent and offspring | 25% | Close inbreeding |
Wright (1922). Generations are counted on the litter's pedigree. Assumes the rest of the pedigree is unrelated and the shared ancestors are not inbred.
What each COI range is called
The calculator sorts a pedigree COI into five bands. The bands name the strategy. None is “good” or “bad”.
A result that lands exactly on a boundary takes the lower band in the calculator, so exactly 25% reads as tight linebreeding there. Breeders still call a full-sibling or parent-to-offspring mating close inbreeding by name. The glossary covers why and when each strategy is used.
Check your registry before a close mating. AKC's registration rules don't restrict them. The UK's Royal Kennel Club won't register a father–daughter, mother–son or brother–sister litter except in exceptional circumstances with its permission, and FCI's breeding guidance says such matings should never be performed.
The bands describe strategy, not health. A 20% litter from health-tested, structurally sound parents is a different proposition from a 5% litter out of untested dogs. What matters is what the line carries.
Why a pedigree COI depends on how far back you go
Compare numbers of the same depth. A deeper pedigree can only find more shared ancestors, so a 3-generation COI never exceeds the 10-generation figure for the same pairing. Our calculator uses 3 generations; many pedigree databases use 5 or 10.
Every pedigree stops somewhere, and the math treats the dogs at that edge as unrelated. In a closed breed they rarely were. That is why a genomic COI, measured from the dog's own DNA, usually reads higher than the pedigree figure.
One study of German Shorthaired Pointers measured both. The 4-generation pedigree COI of 24 dogs averaged 2.3%; the DNA-based figure, from those dogs and ten more, averaged 17%, about seven times higher.
Embark makes the same point about its own test: “Pedigrees can be incomplete or have errors, which can lead to large under-estimation of COI,” while the genetic figure is “a direct measure of realized inbreeding.”
Not every DNA test reports a genomic COI. Embark does. Wisdom Panel reports a genetic-diversity score instead, which it describes as distinct from COI. Both run disease-variant panels. Use the pedigree COI to plan the mating and the DNA results to see what the dogs actually carry.
What studies have measured as inbreeding rises
Linebreeding has costs as well as uses. They are dose-dependent, steeper as COI climbs, and they have been measured in pedigreed dogs:
- Smaller litters. In 93 Golden Retrievers, each 10% rise in DNA-measured inbreeding went with roughly one fewer puppy per litter.
- Exposed recessives. In nearly 2,500 dogs, 90% of the 678 double copies of a known recessive disease variant sat inside the identical stretches of DNA that inbreeding creates. Concentrating a line concentrates whatever it carries.
- More vet care at the breed level. A study that estimated inbreeding for 227 breeds had insurance records for 162 of them, and those showed more non-routine vet care in the more inbred breeds. Body size and head shape counted too: smaller, less inbred breeds were healthier than larger, more inbred ones.
- Lifespan. Across 118 breeds and more than 40,000 dogs, smaller size and more genetic diversity each went with a longer average life and a larger share of deaths from old age.
None of this makes linebreeding wrong. It makes it a tool with a dose, and the breeds we have today show it works when paired with hard selection.
DNA testing handles the recessives a panel can name, and spreading stud use handles the popular-sire problem. The litter-size effect is not something a test removes; it is a cost you weigh against what the pairing fixes.
How inbred each breed already is, measured from DNA
Every breeder starts from the breed's own baseline. The figures below are DNA-based estimates from one study of 227 breeds, where the average across breeds was about 25%.
They are not comparable with a pedigree COI from a calculator, which is why they look high. They describe the breed as a whole, not any one mating.
| Breed | DNA-based estimate | Dogs sampled |
|---|---|---|
| Jack Russell Terrier | 10% | 336 |
| Chihuahua | 11% | 202 |
| Standard Poodle | 15% | 542 |
| Border Collie | 19% | 2,343 |
| Labrador Retriever | 22% | 1,261 |
| French Bulldog | 22% | 1,532 |
| Golden Retriever | 27% | 736 |
| German Shepherd Dog | 29% | 826 |
| Bulldog | 35% | 245 |
| Cavalier King Charles Spaniel | 41% | 455 |
| Pug | 43% | 103 |
| Basenji | 54% | 93 |
Bannasch et al. (2021), Canine Medicine and Genetics, Additional file 1: genotype-based inbreeding estimate (Fadj) per breed, rounded to the nearest percent.
How breeders use the number
Linebreeding works because of selection. Breeders don't pair related dogs at random. They choose sound, type-correct, health-tested dogs to concentrate, and generations of selection remove the ones that don't hold up.
- Run the COI before the mating. It tells you which strategy you are executing: an outcross, a doubling on a grandparent, a tight linebreeding on a producer. Whether the breeding is good depends on the dogs and the goal.
- Choose sound stock. Don't concentrate on a dog you wouldn't want more of.
- DNA panel test. Embark, Wisdom Panel or a breed-specific panel reports each dog's status for the variants it tests. For those variants you know which pairings can and can't produce puppies with two copies, at any COI — and for most recessive conditions two copies is what puts a puppy at risk. A panel says nothing about a variant it doesn't test for, and it does not change the litter-size effect described above, so keep watching what the line produces.
- Track what the line produces. Adjust which dogs you concentrate on real outcomes over generations.
- Outcross for a reason. To bring in a correction the line lacks (a topline, coat density, a temperament trait), outcross to a dog that has it, then linebreed back over the next generation or two to fix it.
Breed-wide outcross programs
A few breeds have gene pools too small to work within. The Norwegian Lundehund went through two bottlenecks, in the 1940s and the 1960s, that each left only five dogs. An outcross program with three Nordic spitz breeds is now under way. That is a breed-level decision, not one for an individual breeder.
The popular sire effect
When one exceptional sire is used heavily, his genetics spread across the next generation of the whole breed. That narrows the gene pool, and it is also how every breed got its top type: the dogs in the ring today trace back to a small number of sires who produced consistently. Each kennel weighs his quality against what heavy use does to the breed.
Disclosure: As an Amazon Associate I earn from qualifying purchases. The shop links here are affiliate links, including Chewy affiliate links — buying through them supports BreedTools and adds nothing to your price, and it never changes what we recommend. Full affiliate disclosure
Get the DNA figure for your own dog
A pedigree COI is the plan. A DNA test shows what the dog actually inherited.
Embark Breed + Health Kit
Reports genomic COIReports a genomic COI from the dog's own DNA alongside screening for 270+ genetic health conditions. Check the condition list against your breed club's protocol: a few breed-specific tests still need a dedicated lab. This is the Breed + Health kit: it gives the tested dog's own COI, not the pair-planning tools in Embark's separate breeder kit.
Related Tools
References & sources
- Wright S. Coefficients of Inbreeding and Relationship. The American Naturalist 1922;56(645):330–338. — The path method behind every pedigree figure here.
- American Kennel Club. The Importance of Extended Family. — Linebreeding as “essential to setting type”; outcrossing used periodically alongside it.
- American Kennel Club. Rules Applying to Registration and Discipline. — Contains no restriction on registering litters from closely related parents.
- The Royal Kennel Club. Litter registration. — Father–daughter, mother–son and brother–sister matings are not accepted, save in exceptional circumstances with permission.
- Fédération Cynologique Internationale. International Breeding Strategies (2010). — “Mating between siblings, mother to son or father to daughter should never be performed.”
- Boccardo A, et al. The German Shorthair Pointer Dog Breed: Genomic Inbreeding and Variability. Animals 2020;10(3):498. — Mean pedigree inbreeding 0.023 on four generations (24 dogs); 0.17 from runs of homozygosity (34 dogs).
- Embark Veterinary. Genetic COI and dog breeding. — How pedigree and genetic COI differ; the Embark for Breeders test includes a COI score.
- Wisdom Panel. Understanding genetic diversity. — Its genetic-diversity measure is distinct from the coefficient of inbreeding.
- Chu ET, et al. Inbreeding depression causes reduced fecundity in Golden Retrievers. Mammalian Genome 2019;30:166–172. — 93 dogs; about one puppy fewer per 10% rise in genomic inbreeding.
- Sams AJ, Boyko AR. Fine-Scale Resolution of Runs of Homozygosity Reveal Patterns of Inbreeding and Substantial Overlap with Recessive Disease Genotypes in Domestic Dogs. G3 2019;9(1):117–123. — 678 homozygous recessive disease genotypes at 29 loci, 90% inside runs of homozygosity.
- Bannasch D, et al. The effect of inbreeding, body size and morphology on health in dog breeds. Canine Medicine and Genetics 2021;8:12. — 227 breeds, mean genotype-based inbreeding 0.249; insurance morbidity data for 162 of them; the per-breed table.
- Kraus C, Snyder-Mackler N, Promislow DEL. How size and genetic diversity shape lifespan across breeds of purebred dogs. GeroScience 2023;45:627–643. — 118 breeds, more than 40,000 dogs.
- Stronen AV, et al. Genetic rescue of an endangered domestic animal through outcrossing with closely related breeds: a case study of the Norwegian Lundehund. PLoS One 2017;12(6):e0177429. — Two bottlenecks that each left five dogs; outcrossing with three Nordic spitz breeds.
This article is for educational purposes and does not replace professional veterinary or genetics advice.
Inbreeding coefficient FAQs
1Is linebreeding bad for dogs?
No — linebreeding is the foundational tool of pedigreed dog breeding and the standard practice in many successful show kennels. Every recognized breed exists because foundation breeders systematically bred related dogs to fix type, and the AKC still calls linebreeding 'essential to setting type.' Breed-average COIs vary enormously — from a few percent in large-population breeds to 25%+ in breeds with small founding stock — and tight-linebred programs that double on a proven producer commonly run well into the double digits. The general public often confuses linebreeding with neglectful breeding; in reality, careful linebreeding paired with DNA testing and honest evaluation produces the predictable, sound dogs that breed standards are built on. The real risks are dose-dependent. DNA testing manages the recessive conditions a panel can name; it doesn't change the smaller litters measured at higher inbreeding, which a breeder weighs as a cost of the pairing rather than chasing a low number for its own sake.
2What is the difference between linebreeding and inbreeding?
Genetically there's no difference — both concentrate the genetics of a common ancestor. 'Linebreeding' is the term breeders use when the common ancestor is further back in the pedigree (great-grandparent doubled, grandparent doubled). 'Inbreeding' is used colloquially for closer pairings like parent-offspring, full sibling, or half sibling. The labels are cultural; the math is the same. Both are legitimate breeding tools used at different times for different reasons.
3What COI do top show breeders actually run?
It varies enormously by breed and by program. Breed-average pedigree COIs run from a few percent in large-population breeds up past 25% in breeds with small founding stock. Within a breed, tight-linebred programs that double on a great producer (a grandparent or great-grandparent appearing on both sides) deliberately run higher to lock in that dog's qualities; foundation kennels rebuilding from a key dog may run higher still. The 'low COI is always better' framing comes from population geneticists worried about breed-wide diversity — that's a real, evidence-backed concern at the breed level — but at the level of an individual mating, what matters most is what's being concentrated, which is why DNA testing beats a target number.
4Why is close inbreeding (sibling, parent-offspring) used at all?
Close inbreeding is a strategic tool. Breeding full siblings or parent to offspring makes it far more likely that a recessive gene carried in the line shows up in the puppies. Serious programs sometimes do one close inbreeding to see what the line is hiding, then select away from whatever it revealed, often with an outcross in the next generation. It is a test with limits: a recessive only shows in a share of the puppies, so a clean litter does not prove the line is clear. Registration rules differ by registry: AKC's registration rules don't restrict these matings, but the UK's Royal Kennel Club won't register puppies from father–daughter, mother–son or brother–sister matings except in exceptional circumstances with its permission, and FCI's breeding guidance says they should never be performed. Check your own registry and breed club before you plan one.
5How many generations should a COI calculation cover?
Compare numbers of the same depth. A deeper pedigree can only find more shared ancestors, so a 3-generation COI never exceeds the 10-generation figure for the same pairing. The BreedTools calculator uses 3 generations; many pedigree databases use 5 or 10. A genomic COI, measured from the dog's own DNA, usually reads higher still, because a pedigree treats its earliest recorded dogs as unrelated. In one study of German Shorthaired Pointers, 24 dogs averaged 2.3% on a 4-generation pedigree, while the DNA-based figure for those dogs and ten more averaged 17%.
6Can two health-tested parents still produce unhealthy puppies if COI is high?
Yes — but the risk is specific. Linebreeding concentrates whatever the line carries. If a recessive condition isn't tested for and is in the line, doubling on the carrier can produce affected puppies. This is why DNA panel testing matters more than COI minimization: a 20% COI litter from DNA-tested, structurally-sound parents who pass everything their breed tests for is a fundamentally different situation from a 5% COI litter produced without any testing. The goal is informed breeding, not low numbers.
7Does COI affect temperament?
Linebreeding concentrates whatever's in the line — including temperament traits. If you linebreed on a stable, sound-tempered dog, you concentrate that. If you linebreed on an anxious or reactive line, you concentrate that too. The 'inbreeding causes nervousness' claim isn't a universal effect; it's a description of what happens when bad temperament is in the line being concentrated.
8Why do some breeds have very high average COIs?
Breeds with small founding populations and closed registries naturally accumulate breed-wide COI over generations. Bulldogs, Pugs, Cavalier King Charles Spaniels, and several toy breeds all trace back to a small number of foundation dogs. This is a breed-level concern (limited gene pool to work with) but it doesn't mean individual breedings within those breeds are wrong — it means breeders have to be especially thoughtful about what's in the lines they're concentrating.
9When does outcrossing make sense?
Outcrossing is a legitimate tool for specific purposes: introducing a structural correction the line lacks, bringing in a missing health clearance, breaking up a fault, or addressing a genuine breed-wide bottleneck. The standard formula in serious programs is: outcross to bring in something specific, then linebreed BACK over the next generations to fix the new trait. Outcross matings without a goal often produce unpredictable type and lose the consistency the program was built on.
10How do DNA tests relate to COI calculations?
DNA panel testing is more important than pedigree COI for actual breeding decisions. Pedigree COI tells you the relationship; DNA tells you what's actually in the dogs. A breeder who DNA-tests everything and breeds informed pairings at 20% COI is doing more for breed health than one running 5% COI on untested random-bred dogs. Embark's breeder test reports a genomic COI alongside its disease panel; Wisdom Panel reports a genetic-diversity score instead, which it describes as distinct from COI. Use both — pedigree COI to understand the relationship, DNA testing to know what's being concentrated.
