When good breeding falls short.

Popular sires, pedigree distortions, and the biological impossibility of inbreeding decline in closed gene pools.

Despite well-intentioned efforts by breeders, genetic inbreeding in Flatcoats is rising. A previous post explains how breeders seeking to minimize inbreeding using pedigree-based COI (coefficient of inbreeding) calculations can still produce litters with high genetic COI.

Embark, a company analyzing canine DNA, recently reported that, over the past year, the average FCR genetic COI has increased from 33% to 33.5%. While it can be difficult to interpret what a shift of this size means at the population level, one point is clear: as the number of FCRs tested by Embark has increased, so has the average genetic COI.

Meanwhile, pedigree inbreeding rates appear to be declining over time, a biological impossibility in a small, closed gene pool. This discrepancy again demonstrates the limitations of pedigrees when calculating genetic and population health metrics.

In 2022, the Royal Kennel Club (RKC) began an extensive analysis of its pedigree database, including FCRs born and registered between 1990 and 2021. One goal was to estimate the UK effective population size (Ne). Most simply, effective population size estimates the size of a gene pool based on how quickly inbreeding (COI) increases over time.

A breed may include thousands of dogs, but because not all dogs breed, and some breed more than others, and some are more inbred than others, Ne helps account for the factors that influence the rate of decline in genetic diversity. A small Ne indicates a population is drawing from a limited genetic pool, meaning diversity will decline more rapidly.

Healthy populations typically have Ne values in the hundreds to thousands (roughly 500–1000+), with higher values associated with long-term genetic stability. Purebred dog benchmarks have already been scaled down; the RKC describes breeds with Ne greater than 100 as “more sustainable.” However, this does not mean genetically healthy—it means the loss of diversity is occurring more slowly, and even a breed with Ne 100 is still functioning as a small, related population that continues to lose genetic diversity over time.

A previous RKC population analysis reported in the 2018 Breed Health Conservation Plan estimated the FCR Ne at around 68, but noted concerns about data quality since imported dog pedigrees often fail to reflect the dog’s true relationship to the UK population. For example, a Swedish import may descend from several UK dogs, but unless those relatives appear within the short import pedigree, that dog can be treated as unrelated and assigned a 0% COI. This artificially lowers calculated inbreeding rates and, in turn, skews Ne estimates.

In the updated analysis, this concern was examined further, and ultimately accounted for, but data quality concerns persisted. So much so, that inbreeding rates were too distorted to reliably calculate Ne. From 1990 to 2008, inbreeding increased as expected in a closed population. After that, however, inbreeding appeared to decline, suggesting an increase in genetic diversity.

While increasing diversity sounds encouraging, it is not biologically possible in closed breeding population. Rather than reflecting real gains in genetic diversity, this apparent inbreeding decline underscores a limitation of pedigree data - specifically, its inability to fully capture how dogs relate to shared ancestors, and especially when pedigrees are short or incomplete.

Just as import pedigrees fail to account for earlier UK ancestry, Flatcoat pedigrees after 1990 fail to fully capture the shared ancestry pre-dating them. As a result, important genetic diversity constriction points, such as limited founders and WWII bottlenecks, are not accounted for in the COI of modern pedigrees.

We have seen many comments suggesting that if breeders focused more on pedigree analysis, rather than using popular or winning sires, the breed’s genetic diversity could be preserved. This perspective oversimplifies the problem.

The RKC has appropriately promoted practices such as reducing popular sire use and increasing the number of dogs from each generation used for breeding. Their analysis supports the idea that breeder efforts to reduce pedigree COI have contributed to recent declines in pedigree-based inbreeding rates. The Flatcoat Conservation Project supports these efforts.

However, it is also clear that FCR inbreeding exceeded healthy thresholds long ago, pedigree-based metrics distort this reality, and breeding for diversity now cannot undo the existing disease burden already pervasive throughout the limited gene pool.

Despite decades of mindful breeding and attention to pedigree relatedness, cancer is (still) killing ~68% of Flatcoats, as reported in the latest Cause of Death Register Report, and similar to mortality stats elsewhere. 

For complex diseases like cancer, genetic diversity is all the more important because it is the only tool available to mitigate the risk of inheriting harmful shared alleles, and also provides greater variation in the building blocks for a dog’s cancer surveillance and suppression mechanisms. Even if we could identify and test for cancer markers, genetic diversity would remain a key benchmark. While recent inbreeding is important for understanding and addressing newly revealed or Mendelian diseases, our breed’s primary health concerns were seeded into the gene pool before modern pedigrees.

This discussion has zeroed in on a lot of UK data points, but that is only because they are monitoring and analyzing the data. This absolutely should not be read as criticism of either the RKC or the UK Flatcoat population. Indeed the opposite - noting these anomalies helps us all identify and account for the gaps and limitations in pedigree-based analyses, and it is encouraging to see honest kennel club efforts towards assessing and improving breeds’ population health. 

Given that European and North American Flatcoat populations are genetically very similar (with a fixation index of approximately 0.03), and all descended from the same limited stock of post-WWII dogs, we can reasonably extrapolate what we learn here and expect comparable patterns across regions. 

Ultimately, these reports serve as a reminder that we are all working toward the same goal: genetically healthy dogs and long-term conservation of the breed.

Pedigree vs. Genetic COI in Flatcoats

The Coefficient of Inbreeding (COI) measures an animal’s homozygosity.  For most practical purposes, you can think of COI as a measurement of genetic “sameyness” or lack of diversity.  We can calculate predicted COIs of prospective matings and we can also aggregate the COI of several individuals to ascertain an average COI for a population.  COI is important because it is one indicator of population health, and it is directly correlated with health outcomes - higher COIs / greater homozygosity leads to poorer health outcomes.

Historically, breeders utilized pedigrees to determine COI, along with a formula that considers the number of common ancestors in a pedigree, the number of generations between the common ancestor and the subject dog, and the COI of the common ancestor if known.

The limitations of pedigree COI are immediately apparent - you’re limited to the data in the pedigree and even 10-generation pedigrees fail to incorporate significant events that have an obvious impact on inbreeding rates.  In Flatcoats, for example, the breed started with a small number of breed founders then suffered a near extinction when the breed was reduced to ~41 dogs during WWII, both of which pre-date a contemporary 10-generation pedigree.  Consequently, dogs can present on paper as being entirely unrelated despite descending from the same handful of ancestors. 

Calculating COI with pedigree tracing back to founders is much more accurate, but is prohibitively processor-intensive and will likely contain gaps that impede accuracy and require that assumptions be made. At it’s most accurate, pedigree COI will still fail to differentiate among siblings, whose actual COIs can vary significantly. It is also important to note that, despite the problems inherent in pedigree-based COIs, pedigree databases can still incorporate powerful tools to optimize diversity outcomes in breeding decisions.

Genetic COIs use an individual dog’s genome to determine COI.  DNA from a saliva sample is run through a microarray where thousands of individual probes examine different locations on the dog’s genome.  This technology has demonstrated reliability and is used extensively for human genetic testing, DNA evidence, and advanced research.  If you want to learn more about how this works, here is a video summarizing the process.

Determining genetic COI is a simple genomic analysis whereby the number of matching alleles are tallied and compared against the total number of loci surveyed.  This project uses Embark, which examines 230K+ markers and provides COI figures for the breed very similar to those offered by whole genome sequence (WGS) analysis.  Any high-resolution SNP profile tool will offer similarly accurate and reliable COIs.

The discrepancy between a Flatcoat’s pedigree and genomic COI is stark and reveals the extraordinary impact of historic bottlenecks on the breed’s population structure.  Here are the COI results of a handful of Flatcoats from varied lineages:

Pedigree COI calcuated by FlatcoatData. Genetic COI determined by Embark.

Embark’s average Flatcoat COI is currently 33% and is increasing as more dogs are tested over time.  To lend a bit of context, here are some COI benchmarks:

• First cousin mating - 6.25%

• Grandparent - grandkid mating - 12.5%

• Parent-child or full sibling mating - 25% 

• Full sibling mating over 3 generations - 50%

• Target ceiling COI of healthy managed populations - 10%

• Three consecutive generations of sibling matings - 50%

As you can see, the homozygosity in Flatcoats has aggregated over time to exceed that of even the closest possible inbreedings, and regardless of how diverse a dog appears on paper.

To learn more about the accuracy and relevance of pedigree, SNP, and WGS COIs in pedigree dogs, this 2016 article by Dreger, et al. provides some illustrative comparative data, while also demonstrating how pedigree COIs can prove beneficial to understanding population architecture precisely because of their inaccuracy.