Let’s talk about relatedness

At BetterBred, we pay very close attention to “relatedness.” This includes a breed’s overall relatedness, and a mating pair’s relatedness, and an individual dog’s relatedness to its breed. Why? Because it is the essence of breeding dogs for specific traits. Too related, and there can be problems. Not related enough, and we lose the attributes we want to see in our breed. Keeping it in the middle is what experienced, conscientious breeders do. We want puppies that are predictable enough to be reliable companions, with temperaments we expect, who can do the work that’s required of them, and who will look the way we want them to look. That requires some of the genes to be similar in all dogs of a specific breed. We also want them healthy enough to have good immune systems, and not to inherit recessive diseases or breed specific complex diseases. That requires them all not to be too related.

The thing about purebred dogs, however, is that all dogs in a single breed are by definition related. HOW related they are is really what breeders must control and there are different ways to measure that.

Often times breeders look at pedigrees to see if there are recent ancestors in common. Sometimes the 5 generation pedigrees of potential breeding pairs will not have any ancestors in common – and breeders will think this means their dogs are unrelated. However 10 generation pedigrees of the same dogs may reveal ancestors in common between to two – sometimes a great many, and 20 generation pedigrees may reveal that the two dogs actually have nearly all the same original ancestors. Thus the two dogs, thought to be unrelated, may actually be quite related.

This fact is at the basis of the alarm for those concerned about genetic diversity in purebred dogs. In some cases, if at the 20th generation, the same few, say 5, dogs are simply repeated over and over, the chances that the eventual descendant, 20 generations on will be highly inbred. They will usually be more inbred than anyone looking at a 5 or 10 generation pedigree might suspect. If, by contrast, the same 50 dogs are repeated over and over in the 20th generation, there’s a much better chance that the eventual descendant will not be overly inbred.

Keeping existing differentiation within the breed is extremely important for the health of the breed, so that there’s less of a chance of fixed negative traits within the breed.

The journey of all the genes through those 20 generation is also a consideration. Were dogs purposefully tightly inbred generation after generation? Or were breeders more moderate about keeping their line tight? Were the dogs bred for aesthetics more than working function? There’s some evidence that working lines retain more diversity. The relatedness of two purebred dogs is a product of both who their ancestors were, and how each generation was chosen along the way.

Breeders often assume there will have been lots of mutations and changes in the DNA in 10 generations. While mutation rates in species differ, for the purposes of purebred dog breeding, there is actually only a negligible amount of mutation in 10 generations. An occasional mutation will appear – but in over 19,000 genes in a single dog, that very low mutation rate means that even 20 generations will not create much variation.

We are stuck, therefore, with the genes that were in all our dogs’ ancestors 20 generations ago or more. This is why all dogs of a single breed (and often a single group!) are all related, and this is the genetic material that purebred breeders work with collectively, whether we know it or not.

But, then – what about 100 generations ago? Are they the same as that? Well, just as with pedigrees there is a different view of relatedness at 5 generations or 10 or 20 generations, there is a different view when considering pre-breed formation ancestors. Are all herding breeds related? Many of them certainly are. Are all mollosoid dogs related? Many are. What about native Asian breeds? Sure they are. If we go back far enough, many breeds are related, because they have similar functions, or because they evolved in the same parts of the world, and they have ancestors in common.

How we think of relatedness, therefore, and what we use to assess genetic relatedness is dependent on how far back we go, and that is dependent on what we need it for.

For those concerned about genetic diversity, the knee-jerk answer to solving perceived diversity loss is to breed any dog to the least related possible dog. The least related dog may in fact be one that has evolved for dozens of generations in completely different parts of the world. Their most recent common ancestor may actually be hundreds of generations in the past. So in fact we do not know what the combination of dogs that have evolved so separately will be. By crossing them, we may in fact lose or diminish many of the traits that were selected for – through natural selection and passive human selection as well as purposeful selection – over a great many generations, many of them prior to the formation of closely related modern breeds. So this is not a practical answer. Diversity for diversity’s sake can be just throwing out eons of beneficial and appropriate evolution.

Certainly a significant loss of diversity is also not a good thing, but to know whether it has actually happened requires appropriate genetic testing using proven methods. It’s wrong to assume all breeds have the same remaining amount of diversity – because they don’t. Each breed has a unique number of ancestors and a unique evolutionary history of its own. This can also be true of lines within a breed. Many of these breed histories are not possible to quantify because records are inaccurate, inadequate, or nonexistent. Complete pedigree databases can be difficult and time consuming to compile and analyze. Luckily we can use genetic testing to give us the current genetic status of any breed.

There are a variety of ways to do that – but when assessing a species that has sub-populations (dogs are the species, breeds are the sub-populations) for the sake of making good breeding choices, it’s important to analyze relatedness first within the context of the species, and then also within the context of the sub-population. This is because the relatedness between all dogs that are within the breed will be far closer than the relatedness between dogs that are within the species, and if you only look at the breed, you may not realize how genetically similar dogs within a breed are. Some breeds, for instance, are so genetically similar within the context of the species, they are all the equivalent of full siblings. As you may imagine, a litter of puppies whose parents were unrelated dogs within the context of the species – let’s say an Akita and a Bloodhound – could be quite dissimilar, one to the next, but they would nevertheless be full siblings. Therefore even a breed made up of all quite genetically similar dogs has some differentiation. That differentiation is what is the genetic diversity in each breed.

How do we know when there’s not enough differentiation? It depends. Some breeds are quite similar – have little variation – but remain healthy. Some are less similar and have many diseases. Every breed is different because every breed started with very different individual ancestors who may or may not have had lurking disease genes, and then have different histories.

We have to base our threshold on outcomes, not some theory or number. Our goal as breeders is to have dogs with predictable appearance, function and temperament who are also healthy. Therefore, how we can gauge whether a breed is too similar is when all the existing families produce the same breed specific diseases. When breeders can no longer find any lines that reliably do not have a common disease or other dysfunctional trait, then there’s simply too much genetic similarity within the breed. That means the genes that cause the problem are “fixed” in the sub-population. The only way to avoid fixed traits is to breed to individuals without the genes for those traits. Sometimes we can still find obscure dogs within a breed that are different enough from the rest of the breed that they are likely not to have the same problematic genes, but sometimes the only option is to outcross to a similar breed without that disease.

It is not remotely necessary – and likely not even advisable – to try to create more “diverse” dogs by pairing dogs up chromosome by chromosome, especially using DNA testing that captures homozygosity or heterozygosity in protein-coding regions, which is true of the major direct to consumer diagnostic and breed identification tests.

Modern purebred dog breeds all are part of a registration system, and most breeders and breed clubs do not ever want to cross breeds. Because of that, no matter how closely related a single breed’s dogs may be within the context of all dogs, what really matters to breeders is how related dogs are within the context of their own breed. Keeping existing differentiation within the breed is extremely important for the health of the breed, so that there’s less of a chance of fixed negative traits within the breed.

The strategy that best preserves existing diversity within the breed is to pair the least related available mates of good quality. And THAT is why BetterBred pays close attention to relatedness.

How do we know this works? We know because it’s similar to any natural isolated population – so many of which have been studied carefully. Animals instinctively preserve genetic diversity through mate choice. Because they tend to have access only to a certain number of potential mates geographically, they cannot always find completely unrelated mates for their species. And so they do what we recommend breeders do to avoid disease and maintain healthy levels of genetic diversity – they instinctively choose the least related available mates. In fact allowing animals to make their own mate choices has recently been shown in zoo populations to increase reproductive success.

It is not remotely necessary – and likely not even advisable – to try to create more “diverse” dogs by pairing dogs up chromosome by chromosome, especially using DNA testing that captures homozygosity or heterozygosity in protein-coding regions, which is true of the major direct to consumer diagnostic and breed identification tests. These regions are the regions most likely to be under selection and are therefore useful for finding known breed specific traits and diseases, as well as unique breed identifying haplotypes. However, by definition, in these areas are both those genes we selected for purposely when breeding healthy dogs with breed type, and the genes we may have selected for accidentally.

By contrast, the test developed by conservation geneticists at UC Davis used very carefully selected microsatellites that were developed for highly specific identification of individuals and relationships for use in forensic analysis. Genetic relatedness calculations are only as good as the dataset, and microsatellites can be particularly useful for genetic diversity assessments because they are considered neutral – in other words, they are not selected for or against. The more loci that are used, and the more alleles per locus, the more accurate the relatedness is. Assessing genetic relatedness using microsatellites remains widely used in conservation genetics, zoo breeding programs, assessments of relatedness and inbreeding in wild populations, and in forensic analysis and population genetics studies because they are accurate and proven. Studies have also definitively shown that a small number of microsatellites are able to show the same genetic relatedness within and between populations as tens of thousands of SNPS.

With over 5100 dogs in the BetterBred database at the time of this post, we also have many thousands of known relationships identified – parents and offspring, full siblings, half siblings, aunts and nieces, uncles and nephews, grandparents, great-grandparents, and cousins of many degrees.

Moreover, the accuracy of any genetic relatedness calculators can easily be checked by comparing to known relationships. With over 5100 dogs in the BetterBred database at the time of this post, we also have many thousands of known relationships identified – parents and offspring, full siblings, half siblings, aunts and nieces, uncles and nephews, grandparents, great-grandparents, and cousins of many degrees. We regularly monitor the accuracy of our estimates by comparing them to known relationships and we are therefore confident in our estimates for breeding purposes and stand by them. Our system therefore is robust and reliable.

The conscientious breeders who use our system, who trust the world’s leading vet school’s lab to produce a reliable canine genetic diversity DNA test, can greatly increase their power to preserve their breeds by using estimates of relatedness of their dogs – both to potential mates and to the rest of the breed – to help select good potential mates. They can also accurately determine which dogs are genetic outliers within the breed, which dogs are most typical for the breed, which dogs are highly inbred and which dogs are highly outbred. Combined with the knowledge of the overall quality and health of the dog itself, understanding genetic relatedness allows breeders to contribute accurately and cooperatively to the long term health and well being of their dogs, lines and breed.