COI Chronicles Part I

The 10% Rule Clarified

A 5-minute read about evaluating COIs correctly.

Coefficient of Inbreeding (COI) calculations have been around longer that anyone reading this post.  COI was developed as a way for breeders to manage risk when breeding “purebred” dogs.  The COI is a mathematical statistic calculated for a breeder to evaluate the downside risk of a breeding.  The risk being “inbreeding depression,” which can cause the loss of vigor and vitality in a breed, as well as the perpetuation of unfavorable physical and behavioral traits.  So, the COI is NOT a measure of the health of a dog nor the health of a potential breeding, as many interpret it, but rather a statistic to be evaluated.  A good breeder will use the COI in conjunction with other quantitative (adherence to the breed standard, hip and elbow ratings, test scores, etc.) and qualitative (cooperation afield, obedience, nose work, prey drive, swimming aptitude, etc.) measures when deciding on a pairing.  Anyone looking for a puppy should do the same when assessing a litter. 

The universal standard for COI is to “keep it below 10%.”  That guideline is pretty well known across breeds, and it was established based on recommendations of canine geneticists.  Unfortunately, not many people know that the “keep it below 10%” COI recommendation is based on a 10-generation pedigree.  Not many breeders nor breed clubs have 10 generation pedigrees, nor are many breeding software packages configured to use 10 generations.  Most software uses either 5, 7 or 8 generations to calculate the COI of any particular dog or to estimate the COI of a planned breeding.  This artificially skews the COI downward, making the risk look much lower than it actually is.  We can illustrate this using the various COIs for Manitou des Terres D’Argos, our Picardy Spaniel foundation sire: 5-Gen:  0.59%    8-Gen:  8.9%     Genetic:  19%

Manny’s 5- and 8-generation COIs are mathematical calculations made with limited pedigree data.  The genetic COI is a more accurate, DNA-based, measure of the level of inbreeding because it evaluates Manny’s actual DNA to identify the level of inbreeding.  As you can see, including 3 more generations of data (5-gen vs 8-gen) increases the COI significantly…~7%.  And the genetic COI is even higher, almost double the 8-generation COI.  So, if you look at a dog with an 8-generation COI and it’s <10% is that good?  Or, if you look at a dog with a genetic COI of 20%, double the 10% rule is that bad?  NEITHER.  Because the 10% rule of thumb is based on a 10-generation pedigree, so it can’t be accurately applied nor properly utilized in either case. 

If a 10-Gen COI is more accurate, why don’t breeder software programs include a couple more generations in calculating COIs?  That answer is partly political and partly logistical.  Politically, a lower COI paints a rosier picture of risk, so using fewer generations can make the risk associated with a planned breeding or any particular dog appear lower than it actually is.  Logistically, adding more generations requires much more front-end work and one helluva lot more software horsepower.  It’s a bitch.  We know, as our Remi has an incomplete 8-generation pedigree in the Picardy Spaniel breed database. 

We had to research and cross reference 3 separate databases to generate an accurate 8-generation pedigree.  We used a spreadsheet to depict her pedigree visually and then we calculated her 8-Gen COI manually.  That part wasn’t too bad. But then we started work on the 9th generation.  We didn’t get very far for 2 reasons:  Adding the 9th generation…approximately 512 more dogs…was a huge task from a research perspective; and, adding that many more dogs to a spreadsheet makes it damn near useless as a visual.  Adding a 10th generation means adding 1,024 additional dogs, a hard pass for even the most dedicated breeders.

Next: COI Part II, How High is Too High?

Ric, Ellen and their 4 Picardy Spaniels live in New Glarus, WI just 25 minutes southwest of Madison.  A lifelong hunter, Ric has trained and hunted Small Munsterlanders, Gordon Setters, and for the past 6 years Picardy Spaniels.  Ellen has an extensive background in animal genetics and a PhD in Reproductive Physiology.  She bred and trained Greater Swiss Mountain Dogs prior to Picardy Spaniels.