All About Outcrossing – Part II

My last post on outcrossing talked about the various definitions of the practice. Of course the reason for outcrossing within a breed is when a breeder’s “line gets too close,” and she is not getting the results she expected. Sometimes that’s in the form of a specific disease or a surprise, sometimes its just a loss of robustness in puppies or too small litters. Traditionally, breeders who linebreed – meaning they breed two relatives – have a sense from a pedigree or their own experience when those relatives are too close, but not always, and it’s a sad thing when they get it wrong and there’s a problem. But until very recently, there wasn’t really a way to know what was really too close. And before you leap to “Just don’t inbreed!” you should know that most long time breeders can tell you stories about the very healthy inbred dogs they’ve known, how they were valuable in their breeding programs because they tended to pass on all the best traits, and how they lived long lives. These stories are true. They do clash with the narrative that all purebred dogs are going to hell in a handbasket due to greedy, ego-driven breeders who just want ribbons and are killing their breeds because the best way to get them is to inbreed. Breeders like that certainly exist, but in tiny fractions of all the breeders.  The vast majority of breeders wouldn’t risk their dogs’ health and happiness knowingly, not ever.

So then what about the part where all the inbreeding is killing all the breeds? The answer is – it depends on the breed, on its history, on the breeder culture, and on whichever dogs were the foundation of the breed, whether that was 20 or 40 or 150 years ago.

Having been on the front lines of assessing my own breed, the first of its kind to have a completed analysis like this,  and its related varieties (Miniature and Toys)  I can tell you that some breeds are in trouble with historical, systemic inbreeding and some are decidedly not. The UC Davis Veterinary Genetics Lab and all our breed communities are fortunate to have the renowned researcher, Dr. Niels Pedersen DVM PhD, still assessing the genetic diversity existing in individual breeds through his groundbreaking work on canine diversity. To date he has peer reviewed and published 4 papers using this method, but he’s actually done at least 19 full analyses of purebred dog populations with the same scrutiny – every one of them different.

A few breeds could absolutely – from a scientific, if not a political viewpoint – benefit from an outcross outside of the breed. Many others can most likely retain their existing diversity indefinitely with good management. I know my preconceived ideas about the state of purebred dogs went out the window a long time ago – I had to see the data on different breeds. And I’m better for it. We usually see the same enlightenment happen with each breed community when their data comes in. Some breeds are managing relatively well with very little diversity and some really have to be careful despite their ample diversity because there’s a genetic bottleneck in the breed, where many dogs are very genetically similar and only some are different. Some have very little diversity, where all the dogs are highly similar genetically and breeders aren’t managing it well – they are inbreeding further not knowing it’s no longer safe, and those breeds make me sad and frustrated. Others have ample diversity and breeders are not making extreme choices, thereby maximizing the retention of that existing diversity, and I want to cheer. But it’s very important not to assume anything until you have some data. 

Ideally, all breeds would be able to survive with proper breed management but for those that may need an outcross, the idea is likely going to cause enormous resistance. No one wants to lose their breed – either to disease or to outcrossing that changes it. Some in the “diversity” camp are so worried about health, they have gone ahead and planned or done outcrosses to very different breeds, without knowing if it’s necessary, and without knowing what the cross will produce. We see people proselytizing on social media for greater difference in outcross for the sake of “diversity,” because they assume that since difference is good, more difference must be better. This is the other side of the breed-solely-for-type coin, where if a long neck is good and super long neck is better. (Actually, no.) 

Then we have occasional traditionalists warning about the dangers of “outbreeding depression,” pointing to real conservation concerns where restored species in the wild don’t always do well with a human managed introduction of new diversity.  But don’t let the name fool you – outbreeding depression is really just a loss of function and survivability when two very different breeds or differently adapted subpopulations are crossed – so it really depends on the cross, and how much we know about each. For example, a Chihuahua and a Great Dane can technically be bred, but those offspring, if they even survive, are not likely to be more viable than either parent,  no matter how inbred they may be.  Obviously, that can be avoided by breeding to a dog with similar functional phenotype, but different enough genetics. What’s different enough? You might be surprised at how genetically different two similar breeds or even varieties can be.

For advocates who are horrified by the idea of an outcross, many are looking for reasons we can’t possibly safely outcross. “You never know what you bring in!”  This actually is often what they say about outcrossing within the breed. But for breeds that are genuinely in need of outcross for health reasons, there is inherently going to be more danger in what’s already “in” than in what might be outside of the breed. Otherwise, why do it? The only reason to outcross is if it’s determined to be beneficial. And there is lots of proof that says it can be, if a breed can’t breed away from a health problem any other way. Our position at BetterBred is that an outcross program should be well planned, club sanctioned and properly documented.

Ironically, there are those who naively post a recent research paper named “Limits to genetic rescue by outcross in pedigree dogs” as “proof” that outcrossing can’t really be done, or is really tricky. This of course is also silly – because most people read the title and not the complex paper, which actually shows how outcrossing is beneficial in most ways, Like many papers of its sort, it is based on a highly specific computer simulation that cannot exist in real life, so readers have to be careful not to extrapolate any conclusions for real world situations. 

The thing about scientific papers is that there are lots of them – because there are lots of ways to look at things. That’s why they are best peer reviewed and published. This particular paper, which is peer reviewed and published, is based on a very small real population and its pedigrees. The computer “outcrosses” to a fictional population and all the genetics of both the real recipient breed and the fictional donor breed are simulated – not based on the actual DNA of the dogs in question.

We are also very grateful to have a huge database of dog data now from many generations of different breeds. This affords us opportunities for looking at changes breed communities have made since they started using the Canine Diversity Test and BetterBred to select less related mates and breedings that produce lower mean kinship. And those results so far are looking quite promising. We will be very excited to share them, as soon as they are ready

One of the problems with computer simulations is that they can only study certain things given narrow parameters, because aspects always interact in real life, so the simpler a simulation can be, the clearer the conclusions can be. From these kinds of simulations, we can sometimes make better informed decisions on certain aspects of breeding. However, no computer simulation yet can effectively tell us how things will go for our breed or population, because in real life there are just so many factors involved in breeding choices that cannot be simulated. Even with clear and simple parameters, each run of any simulation can be different. In real life, we don’t have those parameters. Importantly, this paper does show that proper management in selection is key to the results of breeding – which is exactly what we do at BetterBred. In fact, this is why BetterBred exists.

The inbreeding rate varied widely after outcrossing. In the simulations, mating and survival after an outcross were random, and consequently, reproduction of descendants of the outcross varied across runs. In some runs, descendants were hardly used and this resulted in higher inbreeding rates, and complete elimination of donor alleles. In other runs, descendants were frequently used and inbreeding rates were lower and donor contributions higher. It is thus important to manage the population after an outcross in order to increase the effective population size. There are several options for genetic management, but one of the most effective is the mean kinship method, that is, to exclude animals from breeding that have a higher than average kinship with all other candidates for breeding (Lewis & Windig, 2017). 

Our regular BetterBred members know we use a metric called Average Genetic Relatedness (AGR), which is highly inversely correlated with our proprietary measure, the Outlier Index (OI.)  In other words, when the AGR goes up, the OI goes down proportionately, and vice versa. AGR is the same metric as “mean kinship” mentioned above, except it assesses genetic relatedness based on DNA data, not pedigrees. Study after study has shown DNA to be more accurate than pedigrees. (See this recent paper on naughty penguins for one of many examples of proof!) We rely on that genetic relatedness estimate not only because is it very accurate in predicting known relationships, but there is ample peer-reviewed research that shows minimizing co-ancestry or mean kinship is very important in maximizing retention of genetic diversity.

We are also very grateful to have a huge database of dog data now from many generations of different breeds. This affords us opportunities to look at changes breed communities have made since they started using the Canine Diversity Test and BetterBred to select less related mates and breedings that produce lower mean kinship. And those results so far are looking quite promising. We will be very excited to share them, as soon as they are ready.

Until then, I would urge everyone interested in the health and well being of their chosen breed not to make assumptions about the genetic state of their breed without real analysis using well established methods. Even some small breeds with few founders have done remarkably well and managed their little bit of diversity, while others with ample diversity still suffer from the effects of a severe central bottleneck. Don’t assume your breed is in dire straits and in need of a drastic outcross, and don’t assume your line or your breed is just fine and you’ll never need need any outcrosses. Start with being open minded about discovering what’s really there. You may be pleasantly surprised, or rather alarmed, but fortunately there’s now a constructive way forward no matter what the result.

Natalie Green Tessier