How Can We Use Outcrossing to Restore and Maintain Genetic Diversity?

A summary of “Limits to genetic rescue by outcross in pedigree dogs.” by Windig and Doekes.

Windig JJ, Doekes HP. “Limits to genetic rescue by outcross in pedigree dogs.” J Anim Breed Genet. Jun;135(3):238-248 (2018). doi: 10.1111/jbg.12330.

Link (open-access): https://doi.org/10.1111/jbg.12330 

Summary by: Noah Stetson

This study used a computer simulation to virtually simulate the effects of various outcrossing strategies using data from the Saarloos Wolfdog breed population. The study found that outcrossing almost never introduces harmful genetics into the population. However, it takes a lot of outcrossing to meaningfully increase genetic diversity in a population.

Background

When a population is “inbred” that means the individuals in the population are breeding to individuals they are closely related to. Inbreeding can be measured using a coefficient of inbreeding, which is sometimes shortened as COI. The higher the COI of an individual, the more inbred they are. For example, an individual with a COI of 20% (0.2) is more inbred than an individual with a COI of 5% (0.05). The challenge with inbreeding is that when an undesirable trait pops up, it becomes really difficult to get out of the population because there isn’t much variation in DNA available.

One way breeders help combat inbreeding is by outcrossing, which is breeding an unrelated individual into the population to increase genetic diversity. This could be an individual from a different line within the same breed (for example, a border collie from herding lines bred into a population of border collies from show lines), or it could be an individual from a different breed (for example, a pointer being bred into a population of dalmatians). When outcrossing is successful at restoring genetic diversity, it is called “genetic rescue”.

However, the downside to outcrossing is that the resulting individuals will resemble the rest of their population less. While this can be a good thing (for example, not having a genetic disorder that most of the population has), it could also produce unwanted effects, like not having certain physical or behavioral characteristics that people want in the breed. To bring the population back to resembling its “breed type” after an outcross, breeders may “backcross”, which is when they breed closely related individuals together. For example, a breeder may outcross the population of Dalmatians by crossing a Dalmatian with a Pointer, then backcrossing by breeding one of the crossbreed puppies back to a Dalmatian.

Since people often want to retain most of the physical and behavioral traits in the population they’re outcrossing, an important question to tackle is “How much outcrossing do you need to do to actually make a population more genetically diverse”? Another concern people may have with outcrossing is that by adding an unrelated individual to a population, you could accidentally be bringing in something into the population that you don’t want, like a genetic disease that was not already in the population.

Data & Methods

This study used a computer simulation to simulate the effects of inbreeding, outcrossing, and backcrossing using real pedigree data from 2000-2010 from the Saarloos Wolfhond (aka Saarloos Wolfdog) breed, a breed in the Dutch Kennel Club that started an outcrossing project to counteract high rates of inbreeding. The breedings were simulated (using a computer) to estimate what the genetic diversity would be like after 100 years of breeding, and the simulation was repeated 50 times to see how much variation there was across the different versions. In all the simulations, the breed was simulated without outcrosses for 80 years, and then with outcrosses for 20 years. The researchers simulated two populations: one where the Saarloos Wolfhounds were the donor population, and one where they were the recipient population.

The steps for running the computer simulations were as follows:

1. Population set up with data provided by the user
   1. Population size and age structure are kept constant
2. Animals are mated
   1. The same number of litters are born each year
   2. To produce a litter, a female is chosen at random from the available breeding females and mated with a randomly selected breeding male
   3. Females can have up to 1 litter a year, while males can have multiple litters a year
3. Old animals are “culled” and replaced by young animals
   1. All remaining animals become 1 year older
   2. Culling is determined by age structure and individuals are culled at random

They simulated the following outcross schemes:

• Single outcross, no backcross
  • One litter was sired by a father from the donor population, and the resulting puppies used for breeding were bred as normal (equally likely to be bred to an outcrossed or inbred individual), then more litters were sired from fathers from the donor population, up to 14 times in total
• Single outcross, with backcross
  • 2 or 4 litters were sired by a father from the donor population, and 2 puppies from each litter were used for breeding and bred with an individual from the inbred population
    • Initial outcross followed by 1-4 years of backcrosses
• Repeated outcrosses
  • Outcross for 2 litters
    • Followed by 1 backcross
    • Followed by 2 backcrosses
• Continuous outcross
  • All litters have a 1% chance of being sired from the donor population
  • Averaged out to a single outcross of a single litter repeated 3 times
  • Also simulated with a 5% and 10% chance of being sired by the donor population

Overview of the different breeding strategies that were simulated in this study

In addition to looking at inbreeding, the authors also simulated different alleles (a chunk of DNA associated with a particular trait) from the donor population. They simulated the effects of the following alleles:

• Neutral alleles
• Detrimental alleles, which reduced survival of newborns by 20% with homozygous (when 2 copies of the allele are present)
• Beneficial alleles, which increase newborn survival by 25%
• Lethal alleles, where the dogs die at 3 years of age when homozygous
• Lethal alleles, where puppies die at birth

Results

• The rate of increase of inbreeding in the recipient breed without outcrossing was 2.1% per generation.
• Continuous outcrossing was the only way to get inbreeding rates below 1%.
• The authors suggest that outcross can help “buy time” by “resetting” inbreeding levels to a lower starting point, but should not be used as the only way to manage genetic diversity.
• Outcrossing must be repeated regularly or continuously in order to significantly increase genetic diversity.
• Outcrossing doesn’t work very well when it’s followed by a backcross. Backcrossing just breeds back out the introduced diversity.
  • When 4 backcrosses were used, the inbreeding level actually increased because there was a smaller number of individuals available for pure breeding.
• If outcrossing is followed by backcrossing, it has to be repeated in order to meaningfully increase genetic diversity.
• Why does it take so much outcrossing, even if the first outcross has a COI of 0?
  • It’s because, with each generation of backcrossing, the effect of the outcross is reduced by half.
• Outcrossing was more likely to introduce beneficial alleles than deleterious (harmful) alleles. The beneficial alleles in the simulated donor population were introduced to the simulated population in 100% of the simulations.
• Outcrossing only rarely increased harmful traits (e.g., genetic disorders) into the recipient population. The lethal alleles in the simulated donor population remained completely absent in the recipient population in 98.2% of simulations after doing 3 outcrosses each followed by 2 backcrosses, and remained completely absent in 80.4% of simulations after continuous outcrossing.
  • Note: This is saying that 20% of the simulations had the lethal allele present somewhere in the population, not that 20% of the individual dogs had the lethal allele!
  • In real-world populations that do have the lethal allele present somewhere in the population, breeders can easily identify and breed it out by not breeding the dogs with the disorder, if the allele is dominant. If the allele is recessive, then the disorder won’t show up unless inbreeding is continued, since individuals need two copies of a recessive gene to be affected.

Discussion

This unique study allows us to estimate how different outcross strategies could affect a population. Future research could expand the use of computer-simulated breeding to other breeds and with other breeding strategies. The use of computer simulations could allow breeders to better understand the potential long-term effects of different breeding strategies before it is executed on a real population of animals.

Another thing that’s really cool about this study is that it was produced because of a partnership with the Dutch Saarloos Wolfdog breed club (AVLS). Partnerships between researchers and breeding groups like this one could be a good way to produce more research that helps people make the best breeding decisions.

Something to note about this study is that the population size in the simulation was kept constant. In a real population of dogs, the population could have increased or decreased over time, which could affect the gene pool in ways this study was not able to account for.

Another difference from real life is that in the simulation, the breeding dogs were chosen at random. In real life, breeders don’t choose breeding dogs at random, which could increase or decrease how closely related the breeding pairs are. Additionally, by selecting which dogs to breed, breeders can decrease deleterious/lethal alleles from being introduced from the donor breed into the recipient population relatively easily. If a deleterious allele is dominant, then it’s going to be easy to identify and breed out by not breeding dogs with the disorder. If a deleterious or lethal allele is recessive, then it won’t show up unless you keep inbreeding, since individuals need two copies of a recessive allele to be affected by the disorder.

Also, the effect of outcrossing depends on the size of the recipient population and how related they are to the donor population. For a breed with a larger population than the Saarloos Wolfdog breed, an outcross will have a lower effect than we saw in this simulation, requiring more outcrossing to increase their genetic diversity. This simulation used an unrelated breed as the donor breed, but in real life, donor breeds closely related to the recipient breed may be used.

This study estimated the effects of various outcross strategies in Saarloos Wolfdogs via computer simulations. They found that outcrossing must be performed regularly or continuously to sustainably increase genetic diversity throughout the population. A single outcross (especially when followed by backcrossing) is not enough! This study also found that outcrossing only rarely introduces harmful traits into the population, and more commonly introduces beneficial traits.

This work by the Functional Dog Collaborative is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.