‘Inbreeding’ is possibly the most controversial, and misunderstood, word relating to genetics. In much of the public’s understanding it is conflated with the term ‘incest’ and taken to mean the mating of close (1st degree) relatives, parent/offspring or brother/sister. Yet, some degree of inbreeding is inevitable, particularly so when there is selection. Nevertheless, inbreeding can and does cause serious problems for a population. Therefore, inbreeding must be understood and managed within a population or breed in order to reduce risks.
The fundamental principle of genetics is the resemblance of relatives. That relatives resemble each other implies that there is something shared by relatives (more shared by closer relatives, and less by more distant relatives) that influences these characteristics or traits. These, we now know, are the different variants at genes. If we want to change a trait in a population of domesticated animals, then we breed ‘like-to-like’, i.e. those animals which resemble each other; the biggest, fastest, most biddable etc. So long as these traits are partly influenced by genes, we see a general change in them in the next generation; the offspring resemble their parents and they will on average be bigger, faster, more biddable than the previous generation. However, in breeding from (selecting) those animals that most resemble each other, we are breeding from those that – on average – are more closely related than if mating was random. This, neatly, is also the definition of inbreeding.
Due to its molecular size and way DNA is replicated to produce sex-cells (sperm and eggs), mistakes in copying it are relatively common. These are called mutations and vary in the severity of their consequences from no effect at all to fatal. Because most higher animals have two parents, they carry two variants at each gene – one from each parent. In many cases, having one ‘normal’ variant at a gene is enough to get by. This means that the mutant variant can be unwittingly passed across generations with no outward signs in those individuals carrying it (‘carriers’), and is not ‘picked off’ by natural selection. As a result, every individual will carry several, probably hundreds, of disease causing mutations; research has shown that even humans (a relatively outbred species) carry on average 200-300 mutations that disrupt function, and 50-100 that are a known cause of genetic disease(1). Luckily as we only carry one copy, we are unaffected by the disease. However, when two individuals each carrying a copy of the same mutation mate, there is a 1 in 4 chance of the offspring being affected by the disease. This is what happens in humans with diseases like Cystic Fibrosis, and in dogs with diseases such as Progressive Retinal Atrophy.
Originally all mutations likely arose in a single individual, and have been inherited across countless generations of descendants. Therefore, it means that all individuals carrying a particular mutation are descendants of that individual, and they are related to some degree (although perhaps quite distantly!). So, the problem occurs when two carriers of a particular mutation, i.e. with a common ancestor in whom the mutation arose, mate and risk producing affected offspring. This can occur when the breeding pair are half-siblings, and so have a common parent, or when they are very distantly related descendants from the individual in which the mutation arose, as is the case with Cystic Fibrosis.
Using pedigree (or family tree) information we can calculate the probability of an individual inheriting two identical gene variants from a single ancestral source (including mutations). This is the coefficient of inbreeding (COI). For the example of the offspring of parents who are half-siblings the COI is 0.125 or 12.5%. This means that there is a 12.5%, or 1 in 8 chance, that at each gene the individual has two copies of the same variant which arose from a single ancestor (this is termed identical by descent, or IBD). The COI of 12.5% also means that on average 12.5% of genes will be IBD. Most of these variants will be harmless, some might be due to selection for a particular trait, but some could be a disease causing mutation. Thus, inbreeding risks the emergence of disease due to these mutations.
Think for a moment about the example of offspring from a half-sib mating, where the parents both have a common parent, and compare it to the Cystic Fibrosis mutation which is thought to have first occurred in a person 52,000 years ago. Clearly, even the most rudimentary pedigree records will capture the common ancestry in the first example, and even the most far stretching will not capture it in the second. Our ability to determine the risks from inbreeding, or mating of individuals with common ancestors, depends on how far back we look for those common ancestors. Therefore, the COI also changes depending on how many generations are used in its calculation. It is critical to know the number of generations over which the COI is calculated to be able to gauge its reliability – for example, a COI of 6.5% over 15 generations is higher than one of 5% over 5 generations, but it is also more informative and perhaps a ‘better bet’. There is an easy way to reduce COI, and that is to calculate it over 3 generations – that will give us the bare minimum, showing up the mating of siblings but not cousins. But this gives us a false impression of the risk, and lulls us into a false sense of security.
Unfortunately, the Kennel Club pedigree data for the Keeshonds is patchy. This is likely to be due in part to the Keeshond being a breed with its provenance outside the UK, that was fairly recently recognised by the Kennel Club, and making good use of breeding stock across a range of countries. This will mean that the number of generations of pedigree held by the Kennel Club and used in the calculation of COIs will be low, likely resulting in a severe underestimate of COI. However, the breed is fortunate to have an international pedigree database, which is likely to contain much more complete Keeshond pedigree than does the Kennel Club. Therefore, the COIs generated from this database will be much more reliable since more complete generations of ancestry are included. Given the COI is informative, it would be advisable to make use of the most reliable estimate when seeking the COI of an individual dog, or progeny of a particular mating combination.
While it is important not to underestimate COI by using too few generations of pedigree, a common question is how many generations is enough? The figure below shows how the COI calculated over an increasing number of generations of pedigree relates to the ‘true’ figure (the blue line; COI in this graph is denoted ‘F’). As can be observed, using up to and including 5 generations of pedigree gives a severe underestimate of true COI, underlining what was discussed above. Only when using about 10 generations or more does the COI begin to ‘plateau’, implying that using additional generations adds little or no improvement. The precise point at which the ‘plateau’ occurs will vary between breeds depending on a number of factors, usually occurring between 10 and 20 generations, but the principle is the same. Of course, if less than 10 generations are available, then it doesn’t mean the estimate of COI is useless, any estimate is more useful than none.
However, this shows why it is important to display the number of generations used in calculation alongside the COI. It also shows how COIs calculated from only a few generations of pedigree, as may be the case for Keeshonds on the Kennel Club Mate Select system, may be a severe underestimate. Mate Select provides information on the number of generations used in the calculation of COI, and this is critical to interpretation.
To return to the example of Cystic Fibrosis; Cystic Fibrosis is the most common disease caused by a single variant of a single gene, with about 1 in 2,000-3,000 individuals being affected. This equates to a mutation frequency (similar to prevalence) of about 0.02, i.e. that 2% of all the variants at that gene in the population are the mutant variant. The mutation frequency is quite important as it determines how common the disease will be, given that mating is random with respect to the disease. In dog breeds, as with many domesticated animals, different breeding structures mean that a mutation can become much more widespread in a population than in humans. This is due to popular sires. Popular sires have a lot of offspring, and 2+ generations down the line these popular sires can appear in a lot of pedigrees. This means that the inevitable mutations they carry are spread far and fast. Thus, just as it can be difficult to avoid a particular individual occurring in the pedigree, so there is a higher risk of his mutations emerging as a novel disease in the offspring of two of his descendants. Recent research has estimated the frequencies of some disease causing mutations (for which there are now thankfully DNA tests) as high as 0.282, or 28%; over ten times more common than that for Cystic Fibrosis.
Many physiological processes are quite complex traits, and in contrast to the diseases described above, which are entirely determined by variants at just one gene, are influenced by the environment and multiple (often scores of) genes. Mutations still occur at these genes, but their effect is less drastic – after all, so long as at all the other influential genes an individual carries normal functioning variants, the trait will probably not be affected to any notable degree, i.e. there will be no selective pressure against it. However, as the frequency of these mildly disruptive variants rises, due to a small genepool and possible over use and contribution of a small number of (usually male) individuals, there may be some gradual effect on the traits that these mutations influence. This is known as inbreeding depression, and is commonly observed in fertility traits of mammals, such as failure to conceive, undetected oestrus, and male infertility. Thus, not only do high levels of inbreeding risk the emergence of novel and often catastrophic disease in individuals, but it can also result in a general decline in the ‘fitness’ of the population.
This short article has given a brief overview of why inbreeding causes problems, and the best ways to measure the extent of inbreeding. The COI is a retrospective measure, describing the degree of common ancestry of the offspring of a breeding pair. While we are able to look at COIs of offspring of potential matings ahead of time, and avoid those which are excessively risky, we are constrained by the amount of genetic diversity in the present day population from which breeding animals are drawn.
In cases where inbreeding depression is already threatening the future viability of the breed, possible means to introduce fresh genetics come from outcrosses, either breeding stock from different lines, from breed populations in different countries, or even from different breeds. However, in many cases, whatever the precise current average level of inbreeding, it is possible with effective management to ensure that the remaining genetic diversity in the breed is not eroded too quickly. Monitoring of COIs, and attempting to make use of more distantly related breeding animals is useful, but ultimately the most effective way to prevent rapid inbreeding or erosion of genetic diversity in a breed, is through limiting over-use of popular sires.
(1) https://www.nature.com/articles/nature09534
(2) https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0209864
Dr Tom Lewis, The Kennel Club, March 2019