A new study in press (Dreger et al 2016) reports heterozygosity (inbreeding) for 112 dog breeds, determined several different ways. They computed inbreeding from pedigree data for 5, 10, and all generations in the database. They also determined genomic inbreeding using DNA analysis of SNPs (single nucleotide polymorphisms) as well as whole genome analysis.

I have graphed the data for inbreeding (below) determined from DNA analysis (from their table S2). It is presented both alphabetically by breed (left) and by magnitude of the inbreeding coefficient (right). The lines indicate three levels of inbreeding: green is 6.25% (the result of mating first cousins), yellow is 12.5% (the result of mating half-sibs), and red is 25% (the result of a full-sib cross). For perspective, evidence of inbreeding depression in mammals (e.g., lower fertility, smaller litters, higher puppy mortality, shorter lifespan, etc) first appears at about 5% inbreeding, and the costs of inbreeding (including incresed expression of recessive mutations) usually outweigh the benefits above about 10%. (Read more about the costs and benefits of inbreeding.) Only the Sloughi is less than 6%, and the Chihuahua, Jack Russell Terrier, Tibetan Mastiff, and Spanish Water Dog are less than 12%. There are perhaps 20 breeds with inbreeding of less than 25%, and the rest are 25% or more.

​The highest level of inbreeding by far (> 80%) is for the Norwegian Lundehund. This breed suffers from extremely low fertility and high puppy mortality as well as an often lethal gastrointestinal disorder. As a last effort to save it from extinction, it is now the focus of a cross-breeding program supported by the Norwegian Lundehund club and guided by an international group of scientists who are now reporting the successful production of F2 offspring.

Below are a few of the graphs from the study,which were produced from pedigree data of inbreeding from the beginning of available data to the present. The dashed vertial line shows when the breed was recognized to AKC. (Note that this is not necessarily the date the breed was originally founded.) For each, they depict inbreeding calculated using 5 generations (green), 10 generations (red), and all of the pedigree data (blue). As you would expect (I explain why here), inbreeding from only 5 and 10 generations of data underestimates the inbreeding determined using all generations of data available. In most cases, the magnitude of the error is substantial. (Note that the y-axes are scaled differently on each graph. Apologies for the poor quality of the pictures, where were all that was available in the pre-print of the paper.)

​These data confirm that inbreeding of most purebred breeds is extremely high, with all but a handful of values exceeding the level of inbreeding produced by mating full siblings from unrelated parents.

Dreger DL, M Rimbault, BW Davis, A Bhatnagar, HG Parker, & EA Ostrander. 2016. Whole genome sequence, SNP chips and pedigree structure: building demographic profiles in domestic dog breeds to optimize genetic trait mapping. http://dmm.biologists.org/lookup/doi/10.1242/dmm.027037

Individual Certification
Dogs that are DNA tested and genotyped through ICB will have access to the ICB Breeder Tool, and they can also receive Genetic Diversity Certification. This indicates that the inbreeding coefficient of that dog was determined directly from DNA analysis by ICB.

In addition to diversity certification, dogs with inbreeding coefficients less than 10% will be provided with a watermarked and coded emblem that can be displayed on the website of the breeder or owner. The gold Diversity Certified emblem will designate an inbreeding coefficient of 5% or less, and the blue Diversity Certified emblem is for inbreeding between 5% and 10%. This fee for certification is $15 and is valid for the lifetime of the dog.

Litter Certification
We are also providing a way to certify litters without genotyping each puppy. This certification is called "Diversity Bred". The two parents are genotyped and the average inbreeding coefficient of the puppies is estimated from the genomic kinship coefficient.

​

Check out
ICB's online courses
​
*******************

Join our Facebook Group
ICB Breeding for the Future
...the science of dog breeding
*******************

Visit our Facebook Page
ICB Institute of Canine Biology
...the latest canine news and research
​

The last decade has seen a dramatic shift in the way we approach dealing with genetic disorders in dogs. Modern genetic technology now allows us to identify single mutations that cause disorders relatively easily. A DNA test is made available, added to the list of routine genetic screens run by the responsible breeders, and the dog is declared to be "health tested". 

As I have pointed out elsewhere, DNA tests screen for mutations, not health. We know with certainty that in every dog and every breed there are many more mutations we don't know about yet, simply because they haven't been manifested as a disease in a dog. A dog can be quite healthy, but we can be sure it carries in its genes dozens of genetic errors that are benign as singletons, but inherited as a pair can produce a serious problem.

The author argues for a shift in perspective that has profound implications for how we approach the genetic basis of disease in humans, and certainly also in dogs. Our approach to genetic testing is based on the presumption that we can produce healthy dogs by eliminating the specific genes that cause them, at least for monogenic disorders. ​The author suggests otherwise.

Untangling complex relationships affecting expression of mutations thought to be Mendelian is going to be difficult at best. There are futher problems when flawed test protocols turn out to be the source of discordent findings, as was recently shown to be the case for the testing for a mutation associated with degenerative myelopathy (Turba et al 2016). (A primer mis-match was found to be the cause of incorrect genotyping for the SOD1 genes.)

The notion that we can solve the problems of canine genetic diseases, even those thought to be caused by simple monogenic mutations, is misguided. The health status of a dog is not the sum of its known mutations. The physiological and biochemical process behind every trait and function are complex, and the genes and other factors that influence their expression are even more so.

​Breeding against specific disorders is not the path to achieving healthy dogs. We must take a larger, more wholistic view, recognizing that we don't understand all of the genetic and environmental factors that might interact in extremely complex ways. If we want to achieve health, we need to breed for the health of the whole dog.

​Take a few minutes and read this thought-provoking paper. It might make you see the path towards truly healthy dogs in a very different way. You can download a copy below.

​

Katsanis N. 2016. The continuum of causality in human genetic disorders. Genome Biology 17:233. DOI: 10.1186/s13059-016-1107-9

Turba ME, R Loechel, E Rombola, GGandini, & F Gentilini. 2016. Evidence of a genomic insertion in intron of SOD1 causing allelic drop-out during routine diagnostic testing for canine degenerative myelopathy. Anim Gen; doi: 10.1111/age.12525

About 70% of the genetic disorders in dogs are caused by single, recessive mutations. In most cases, a dog with only one copy of the mutation will be unaffected. This means that breeders can avoid these disorders entirely with DNA tests and appropriate genetic management strategies like reducing the level of inbreeding and preventing the spread of mutations by popular sires. 

But what about the other 30% of problems for which the genetics are complex or unknown? There are many of these, like heart disease, cancer, epilepsy, hip dysplasia. These disorders are extremely difficult to manage because we don't have good tools to assess risk or predict whether a dog is likely to be affected.

This is a problem we specifically wanted to address in the design of the new ICB Breeder Tool.

For genetic conditions, risk is greater for related dogs. You can get a general idea of the relatedness of dogs using pedigree information, but every puppy in a litter inherits a different mix of parental genes. Pedigree analysis can provide an estimate of risk, but it can't tell you if a particular dog actually inherited the deleterious alleles for a genetic disorder. If we don't know which genes are causing the problem, what can we do?

The ICB Breeder Tool provides a way to determine the risk of a genetic disorder without knowing which genes are involved. Here's how it works.

The ICB Breeder tool takes the DNA information from a high-resolution analysis and compares each dog with the others in the population at each of thousands of markers. Dogs that are related will tend to share more of the same markers, and the more closely they're related the more similar they will be. This information is then used by the ICB Breeder Tool to construct a diagram that clusters together the dogs that are most genetically similar, and therefore most closely related. This treelike diagram is called a dendrogram. (Learn more about how to interpret a dendrogram.) 

This is a dendrogram that shows the genetic relationships among 11 Chinook dogs. Each of the vertical lines across the bottom is an individual dog.

The analysis has revealed two genetic groups: one on the left with three dogs in red and two in green, and on the right there is a group with two blue, two magenta, and two yellow dogs. The dogs that are most closely related are connected by the fewest, shortest lines. So there is one red dog that isn't as closely related to the other two as they are to each other, but it's more related to the other red dogs than the green ones. This diagram tells us about the actual genetic similarity or relatedness of the dogs in this group.

Now, if we are trying to manage a complex genetic disorder, we can use this dendrogram to help identify where the deleterious genes are in the population even though we don't know what they are.

They didn't need to know which genes are responsible for elbow dysplasia. They didn't need to do a research study. They were able to use the information they had about genetic relationships to pinpoint the location of the problem in their breeding stock and reduce the risk of problems in the puppies they produce going forward.

This is a great demonstration of how you can use information about genetic information to effectively manage a disorder with complex or unknown genetics.
​
This kind of analysis can be used for any genetic disorder or trait, without needing to know anything about the mode of inheritance or the genes involved. It can also be used to identify emerging problems that are appearing in certain lines. For problems like hip dysplasia, cardiomyopathy, autoimmune disorders, cancer, and even temperament, this simple analysis can identify the groups of dogs most likely to be at risk. There is no need to do research on the disease or identify the genes involved.

The ICB Breeder Tool uses the data from a high-resolution DNA analysis to construct the dendrogram of genetic relationships among dogs. This can then be used to provide accurate information about the distribution and incidence of a genetic issue in a population of dogs.

​The ICB Breeder Tool is available now to breeders. You can read more about the other ways it can help you reduce the risk of genetic disorders and improve the health of your next litter of puppies.

Genomic analysis is going to revolutionize dog breeding!

​REFERENCE
Ubbink GJ, J van de Broek, HAW Hazewinkel, & J Rothuizen. 1998. Cluster analysis of the genetic heterogeneity and disease distributions in purebred dog populations. Vet Rec Feb 28, 1998.

Check out
ICB's online courses
​
*******************

Join our Facebook Group
ICB Breeding for the Future
...the science of dog breeding
*******************

Visit our Facebook Page
ICB Institute of Canine Biology
...the latest canine news and research
​

The reason breeding is both art and science is because we can only know about part of the science. The basics of inheritance are straightforward, but beyond that things get really complicated. Plus, we will never remove the element of chance from the genetics of breeding because when the alleles get passed out to the next generation, who gets what is a matter of probability. But there is lots about breeding dogs that we can predict, if not perfectly then at least in a general way, and the more of those things we can get right the better. 

I'm going to tell you about a few of the things becoming available to breeders in the new Breeder Tool being developed by ICB. Some are ready to use now, some are in progress, and there are more on the development list. Together, they will provide a package of tools that can help breeders identify existing problems and avoid creating new ones.

​The ICB Breeder Tool is an online, interactive genetic management system that incorporates pedigree, DNA, and health information in a single, integrated framework. It is designed to address the genetics of both your next litter and the breed as a whole, so you can meet your goals as a breeder at the same time as you ensure that the health of the gene pool for the future.

Below I will give you several examples of what the Breeder Tool does, after some basic information about the DNA analysis.

The ICB Breeder Tool is based on analysis of about 230,000 markers distributed evenly over all of the chromosomes, including the mitochondrial DNA (which is inherited from the mother) and the Y chromosome (which is inherited from the father). There are 38 autosomal chromosomes in the dog, plus the X and Y sex chromosomes. That means that there are on average more than 5,000 markers on each chromosome.

  

Homozygosity & Inbreeding
​
About 70% of the genetic problems in dogs are caused by recessive mutations.

​A dog that inherits two copies of a broken gene will be affected. The unique features of every breed are maintained by breeding related dogs, but related dogs also are likely to share mutations and this is what causes the problems. Managing inbreeding is the only way to effectively reduce and manage genetic disorders caused by recessive mutations, so knowing the genetic relatedness of dogs you want to breed is essential.

You can compute the coefficient of inbreeding (COI) from a pedigree database. This will give you the probability of inheriting two copies of the same allele (or mutation) that originates in an ancestor on both sides of the pedigree. If you have an error-free pedigree database that goes back to founders, you can get very good estimates for the inbreeding coefficient. But most databases are less than perfect and many are incomplete.

We can use DNA to get two important pieces of information:

• the amount of genetic heterozygosity of the dog
• the genetic similarity of two or more dogs (e.g., comparing a bitch to a potential sire, or the puppies in a litter with each other.​
  

​
The two nucleotides at each SNP (paternal and maternal) are identified, and the fraction of the genome that is homozygous and heterozygous is reported as "Heterozygosity", which ranges from 1 (all SNPs are heterozygous) to 0 (none heterozygous). This is calculated by dividing the number of heterozygous SNPs to the total number of markers evaluated.

Evaluating Breeding Options

This is where the rubber meets the road. The more you know about the genetics of the dogs you consider for sire, the better you will be able to predict what you will get in your next litter. DNA testing can now provide information about some specific genes of importance for both health and physical phenotype. But we have tests for only a fraction of the known genetic disorders in dogs, and there are also the issues of fertility, lifespan, temperament, and other traits for which many genes are involved. We need a management strategy for those as well.

The evidence is overwhelming that levels of inbreeding higher than about 5-10% compromise many aspects of health, and especially the traits related to reproduction and what the old-time breeders called "vigor"; health and vitality might be what we would call it today. There are many signs of the negative effects of inbreeding depression in today's purebred dogs such as missed breedings, resorbed puppies, small litters, puppies that fail to thrive, and bitches without adequate milk or that don't care properly for the pups. Of course, inbreeding also increases the expression of diseases caused by recessive mutations. For all of these reasons, the single most effective thing breeders can do to improve the health of purebred dogs is to reduce the level of inbreeding.

Most breeders already use pedigree databases to determine the coefficient of inbreeding when they plan a litter, and this remains the single most cost-effective way to manage genetic risk. But accuracy depends on the quality of the database, and unfortunately most are not as good as they should be. Also, COI is a probability, a prediction of the average level of inbreeding to expect in the offspring of a particular mating. But every puppy gets a different mix of genes so each will have a different amount of actual homozygosity, and in a litter of 6 or 10 puppies there might be none that have the predicted COI. Some will be higher, some will be lower; it's simply a matter of chance. DNA can offer much better information about the genetic composition of a dog, and we can use this information to provide much better information about genetic similarity.

The ICB Breeder Tool can determine the genetic similarity of two animals by directly comparing the DNA. Dogs that are less similar genetically will produce offspring with lower inbreeding and lower risk of genetic disease. The breeder could therefore use this information when making breeding decisions.

Let's say you have a bitch to breed and have chosen four potential sires, any of which would be acceptable in terms of temperament, breed type, and health. Now the decision is to determine which sire will be the best match for the bitch in terms of managing inbreeding in the offspring.

Before deciding which sire to use, there are other things you might want to consider in addition to the relationship coefficient. For instance, sire #4 might be a carrier for a mutation also carried by the bitch, which of course would risk producing carrier and affected puppies. You might also want to compare genetic similarity of the regions of the chromosomes where the genes for the immune system are located (e.g., the dog leukocyte antigens, DLA, on chromosome 12). ​

There are other ways you can use this tool. You might use it to select which puppies to keep for breeding based on what they can contribute genetically to your breeding program. You select a boy puppy that will complement a bitch you would like to breed to, or you might want to keep two pups from this litter and would like to pick the two that are least similar to each other. The RC of a pair of dogs is also about 2X the average inbreeding coefficient of a hypothetical litter of puppies, so sire #3 would produce very high inbreeding (24%, about the same as a full-sib cross) and sires #1 and 3 would produce puppies with very low inbreeding (3.5% and 2%, respectively). 

How do you find a dog to breed to that will bring some fresh blood to your line? About the best you can do is sit down with a pile of pedigrees and make lists, or if you're very lucky and have a good pedigree database you can produce some inbreeding coefficients from potential test matings. There are better, easier ways to do this.

The ICB Breeder Tool can compare the DNA data for many dogs to determine genetic relationships among them and display this information in a variety of ways.

Below are several "dendrograms" - tree pictures - that cluster dogs together based on their genetic similarity. Dogs connected by shorter vertical branches are more similar to each other, and dogs connected by longer branches, or in different clusters, are less related. The dendrogram is a picture of the "genetic structure" of your population by identifying subpopulations or groups of dogs that are more closely related to each other than they are to dog in other clusters. If you're looking for an outcross, consider dogs in other clusters, and the further away they are the less related they are. ​

In this dendrogram for American Cocker Spaniels, there are two clear groupings, of the green dogs on the left and the red dogs on the right. Each vertical line is a different dog. Short vertical lines might indicate litter mates or half-sibs. I don't know who the dogs are, but the data probably includes dogs from both the US and UK, and it is possible that these grouping represent genetic differences that reflect geography. (You can learn more about how to interpret dendrograms HERE.)

Learn more about using genetic relatedness and dendrograms to manage genetic disorders when you don't know anything about the cause.

These diagrams cluster the animals by genetic similarity and, together with information about a trait or other characteristic of the dogs, it can reveal subpopulations or clusters and show how they are related. Here you can see clusters of dogs breed for different purposes (top), and below you an identify the relationships among dogs with different health issues like cancer, epilepsy, or heart disease. 

The ICB Breeder Tool provides the most powerful genetic analyses ever available to dog breeders, and there are more features coming (including estimated breeding values!). We are focused on producing tools that breeders can use NOW, so you can start improving the health of your puppies with the very next litter.

​Check out the list of available DNA tests and a summary of the analyses included in the report for each dog.

The ICB Breeder Tool is in beta testing and is available now at a great price. 

Runs of homozygosity
​
Are you a Power User? If you are, read on!

We're going to show you something that seems a little crazy and a lot more complicated than it really is. But it's really, really cool, so don't let your eyes glaze over. I promise, it will be worth it!

Inbreeding results in homozygosity, and in more inbred animals regions of homozygosity tend to form blocks that get longer and longer. These are called "runs of homozygosity" (ROH). 

Because a DNA analysis can tell us not only the amount of homozygosity in an animal's genome, it can also show us where it is, it is possible to produce a map of the chromosomes that displays these regions of homozygosity. We can line up the chromosomes end to end, from 1 to 38, and color code each SNP marker as blue for homozygous or red for heterozygous.

​This is the ROH scan for a Whippet. The chromosome numbers are indicated across the top, and the strip below indicates the location of markers that are homozygous in blue, and red indicates heterozygous markers.

You can see areas where there are regions of high homozygosity as blocks of blue, such as on chromosomes 3 and 4.

We can zoom in on a scan to see more detail. This is a set of scans for three animals with the chromosome numbers across the top indicating that this is chromosome 7 with some of 6 and 8 on either side.

You can clearly see that the first dog has long stretches of high homozygosity, while the regions of homozogity of the other two dogs are much smaller and more evenly distributed over the chromosomes. 

There are some cool things we can do with these ROH charts. 

First, because they document homozygosity, we can use the total length of the homozygous blocks to compute the actual level of inbreeding. We simply add up the lengths of the blocks and divide by the total length of the genome, and multiply by 100 to express it as a percentage. So for the dog above using the scan for the shortest blocks (70kb), the inbreeding level for this dog is 18.6%.

​Another useful thing we can do is compare the ROH scans for multiple dogs at a time. Below is a panel of scans for 12 Bernese Mountain dogs, all scanned for runs of homozygosity longer than 1000 kb, over all 38 chromosomes.

How can a breeder use this information?

​Let's say that you're contemplating using several different sires with similar levels of inbreeding. By comparing the patterns of homozygosity in a bitch and potential sire, you can choose the dog that reduces the areas of shared homozogysity and will produce less homozygosity in the puppies.

Runs of homozygosity also provide a really interesting way to compare the inbreeding in different breeds of dogs. Below are scans of multiple dogs (all at the same resolution of ROH>1,000 kb) for village dogs from India, mixed breed dogs, Jack Russell Terriers, Dobermans, and Chinook. The differences among these groups are striking. Blocks of homozygosity increase in length with inbreeding and they also tend to have high linkage, which means that entire blocks tend to be inherited together. Obviously, selection for particular traits gets harder and harder as these blocks of homozygosity increase in length, and of course the amount of genetic variation is low in these areas, both of which make things more difficult for the breeder. Outcrossing reduces homozygosity and tends to break up these blocks of homozygosity, which is why the first generation from an outcross might not be exactly what you're looking for, but improvement and refinement is much easier.

There is more information and examples of Runs of Homozygosity in our two-part tutorial HERE.