There is ongoing discussion about using a low number of generations when calculating coefficient of inbreeding and whether that gives an accurate assessment of the measure of inbreeding. Some people will claim that the genetic bottlenecks in the breed as well as the low number of founders means all Irish Wolfhounds are highly inbred. It’s a known fact that using a higher number of generations when calculating inbreeding means an increase in the coefficient of inbreeding.

You will always get inbreeding

Average inbreeding for 30 generations is 32.29%

No matter how careful you are to use differing bloodlines in your breeding program, pure mathematics indicate that it’s impossible to get individuals that don’t have related parents. That’s the same in all species, including humans. It’s just a matter of how many generations you use to calculate the levels of relatedness. For Irish Wolfhounds, calculating 30 generations means there are more than a billion possible ancestors (1,073,741,822 to be precise). Given that the number of Irish Wolfhounds that have lived in that period is almost certainly less than 200,000, inbreeding has to occur in all animals. Adding to that, more than half of the Irish Wolfhounds that ever lived were born in the last 25 years. Fluctuations in population size will almost always lead to increased inbreeding if you calculate enough generations. The graph below shows how the population has developed (number of hounds born per year) over time:

Development of population over time

Steep increase from 10 to 20 generations

We have calculated the coefficient of inbreeding (COI) for all Irish Wolfhounds in the database from 2 through 30 generations. We have then looked at the dogs being born in the last ten years to see what level of inbreeding we are seeing when calculating back as far as the software we used can give meaningful results. For most dogs we do have complete pedigrees going back to the breed’s founding animals.

When doing this for the Irish Wolfhound, we see quite a steep increase in coefficient of inbreeding from 10 to 20 generations. The average COI when calculated for 10 generations is 6.5%, while the average calculated for 20 generations is 27.98%. After 20 generations, inbreeding levels flatten out, and the average for 30 generations is 32.29%, while the increase from 25 to 30 generations is very low. The graph below shows you how inbreeding develops over time:

Development of inbreeding

Low inbreeding matters

An interesting finding when doing these calculations is that low COI levels in low numbers of generations usually also correlates with a lower coefficient of inbreeding  when we calculate back to the founders. This observation led us to do the same calculations on two seperate populations of wolfhounds; those dogs that have a relatively low COI (less than 8%) at 10 generations and those that had a higher than 8% COI at 10 generations. The average COI’s of both populations can be seen in the graph above compared to the average of the entire population.

The results clearly show a direct correlation between a high 5 generation COI and a high 30 generation COI. The average “less inbred” dog had a COI of 1.20% at 5 generations, while the average “more inbred” dog had a COI of 9.20%. If we move on to 30 generations, the “less inbred” average is 30.16% while the “more inbred” has an average COI of 38.03%. As you can see, the difference between the average COI of these populations at both 5 generation and 30 generations is steady at 8 percentage points. In practice this means that a mating with a low COI at a practical number of generations (in, that would be 10) also has a lower than average COI when we consider higher numbers of generations. Therefore, it’s meaningful to check COI using the trial mating feature in when you plan a breeding.

Inbreeding recommendations

The UK Kennel Club recommends breedings to have a lower COI than the average for the breed. The following table gives average values for the number of generations we can calculate in based on data from the last ten years: