Australian Shepherd Health & Genetics Institute

Australian Shepherd Health & Genetics Institute

Generic selectors
Exact matches only
Search in title
Search in content


The Breeders Road Map

by C.A. Sharp
1449228116_Czech RepublicFirst published in Double Helix Network News, Spring 2004, Rev. May 2013 France-flag


Photo By: Larry Green
Photo By: Larry Green

Road maps tell you where you’ve been and where you are going.  If you are breeding dogs, pedigrees are your roadmap.  They don’t just tell you what has been.  Used properly, they can give you a good idea of where you need to go.  This doesn’t mean that your goal is to make something that looks pretty on paper.  You’re breeding dogs, not documents but those documents contain a wealth of data and can point you toward additional information.  Analysis of pedigrees and supporting information will aid you in making informed breeding decisions.

Pedigreed Dogs

The use of written pedigrees is so intrinsic to the breeding of purebred dogs that the general public views them as synonymous.  Pedigree = purebred.  But a pedigree isn’t just part of the documentation you send along with a pup when it goes to its new home. 

The standard pedigree format, sometimes called a horizontal pedigree, is a listing of recent ancestors ranked by generation.  Along with the names of the ancestors you may or may not find additional information on such things as registration, appearance or date of birth.  Most printed pedigrees show only three to five generations.  The deeper the pedigree and the more supporting information it contains, the more useful it is as a research tool.  However there are limits to what can be put on a single piece of paper.  A great deal of vital information about the listed dogs cannot be found on a printed pedigree.

Vertical pedigrees show not only your dog, its parents and grandparents, but all of their full siblings.  It may also contain information on things like hip status and can be useful in determining whether polygenic traits like HD may run in the family.  Like the horizontal pedigree, the information listed is limited by the size of a piece of paper.

There is another type of pedigree, often called a genealogy chart, flows the opposite direction—from one or more ancestors down through their many descendants.  These are commonly presented as a chart with squares and/or circles representing individuals and lines indicating mates and offspring.  The number of generations shown depends on the reason for constructing the chart.  Genealogy charts may also contain a limited amount of information about the individuals listed.  Researchers use this type of pedigree to demonstrate patterns of descent for hereditary traits.  A breeder might use it to show important descendants of one of her dogs.  If you are a visual person, sketching out a pedigree descent chart and noting what you know about individuals in it may help you understand how a trait is flowing through your line or indicate how significantly a particular dog has impacted breeding program.

For the most part, we rely on the classic listing of the dog, its parents, grandparents, and so on.  Generating pedigrees once required hours of tedious typing or hand copying.  Thanks to computers, kennel management and pedigree software are readily available.  Paper pedigrees can be printed as needed or posted on websites, sometimes customized to include additional information on the dogs, kennel logos, or photographs.  Considerably more data can be stored on each individual dog’s record than was ever possible on a paper pedigree.

Keeping Track of Business

Whether you use paper records or a computer, you should record as much information as possible about all your dogs and as many of their relatives as possible.  Also, record important facts about dogs that you may use in your breeding program.  Don’t neglect their relatives, either.  The better kennel software will calculate the coefficient of inbreeding (COI.)   You can use COI to monitor the level of inbreeding in your dogs or proposed matings.

Beyond the standard descriptive data and titles, note the strengths and weaknesses of each animal.  If it has been bred, what do you know about its offspring?  Have any of its relatives ever had an inherited disease or disqualifying fault?  With this kind of information readily available, you may be able to focus on the best additions to your kennel or prepare a short list of potential studs for your bitch without ever leaving home or picking up a telephone. 

For example, if your bitch’s weakest point is straight stifles, you could pull up data on dogs you think might be complementary.  What do your notes say about their stifle angulation?  Remember to check their offspring—knowing the dog has consistently produced proper angulation will be a further point in his favor.  Look at his siblings, parents, cousins, and other relatives.  The more of them that have correct stifles, the more likely the dog is to produce them.  If your bitch has been bred before, records on her offspring will indicate how frequently she throws her faulty stifles and whether any prior crosses resulted in improvement.  If straight stifles are something you have been dealing with for several generations, look back at records from prior breedings.  Which crosses resulted in consistently correct angulation?  You can determine how likely you are to get a particular trait so long as you have enough pedigree data.  This can be done for any trait, good or bad.  It’s no guarantee on the outcome, of course, but the point of the exercise is to map out the route that is most likely to get you to your desired destination.

Number Crunching

If a trait is due to a single gene you can calculate the probability that a dog will have the trait, based on which individuals in the pedigree have exhibited or produced it and what is known about how the gene is passed. 

Liver is recessive to black.  If your bitch is liver and you want to breed her to a black dog, you know that all the pups will at least carry liver because your bitch only has liver versions of the gene.  The probability that they at least will be carriers is 100%.  But what is the probability that you will get liver pups? 

If you don’t know whether or not the sire is carrying liver, you need to look at the pedigree.  If one of his parents was liver, he does.  There is a 50% probability that a given pup will be liver.  If both his parents were black, but the dog has a liver littermate, both his parents are carriers and there is a two in three chance he is carrying liver himself and a one in three that you will get a liver pup if you breed him to your bitch, or 33.3% probability. 

What if you don’t know the mode of inheritance for a trait or it is polygenic?  The best way to calculate whether a given cross will produce a trait of unknown or complex inheritance is with Best Linear Unbiased Prediction {BLUP) analysis.  BLUP has been used successfully to maintain high levels of marketability in livestock.  It enables breeders to determine Estimated Breeding Values (EBVs) for herds and flocks as well as individuals, compensating for management and other environmental differences.  The EBVs of one individual or group can be meaningfully compared to those of others.  This helps breeders select stock that will best meet their goals, whether they are breeding commercial production animals or bloodstock.

BLUP can be applied equally well to dog breeding:  It is used by Canine Companions for Independence and Seeing Eye, as well as some European breed organizations.  At present none of the commercially available kennel pedigree software will perform BLUP analysis.  The math and statistical analysis required to do it by hand are enough to make ordinary mortals tremble.  It also requires a data set far more comprehensive than most breeders will be able to pull together.

A more practical pedigree analysis technique for the average dog breeder is a modification of percentage of ancestry.   Percentage of ancestry, sometimes called percentage of blood, is used to determine how much an individual ancestor contributed to the pedigree.   Each parent will have given your dog 50% of its genes.  Statistically, each grandparent will have given 25%, great grandparents 12.5% and so on.   There comes a point when the probability is very low that your dog will have inherited any significant number of genes from that particular ancestor unless the ancestor appears in the pedigree multiple times.  If a single dog is four times a great grandparent, he will have contributed about as many of his genes as did either of the actual parents—50%. 

   Oso is a grandsire and a great grandsire:  25 + 12.5 = 37.5% ancestry.
   Lady is great grand dam twice and great-great grand dam three times:
      (2 x 12.5) + (3 x 6.25) =  43.75% ancestry

Oso falls between a grandparent and a parent in his potential contribution to the pedigree.  Lady is very nearly at the level of a parent.  I say “potential” because we cannot know exactly which genes came down through the intervening generations.  It may be everything the individual passed to its immediate descendant or it may be nothing.  Dogs have around 20-30,000 genes, so in most cases the actual number of genes passed down will hover somewhere around the percentage of ancestry. 

Percentage of ancestry cannot exceed 50%.  No bitch or dog could be behind more than half of the possible lines of descent in a pedigree.  If you come up with a number greater than 50, you have made an error.  Percentage of ancestry allows you to determine how much individual ancestors may have contributed to your dog’s genes.  With minor modifications, the same technique and can be used to determine how likely you are to get a particular trait. 

In-Depth Analysis

To determine the risk you will get particular traits, review the pedigree for dogs that have had or produced the trait.   Except in the case of X-linked or single gene dominant traits, both parents of any individual exhibiting the trait will be carriers of genes for that trait.  I usually go a step farther and note grandparents of individuals that had the trait.  With polygenic traits, it is highly probable that the grandparents also carry genes for it.  It’s also very likely that you won’t have full information, especially if the trait is one people don’t want to admit their dogs have.  However, you might know about one or more affected grandpups.  I refer to grandparents of affected dogs as “suspect” carriers.  Suspects should not be given the same weight as dogs that actually exhibited the trait or produced it, but including them in your analysis is as a technique for incorporating the breath of pedigree necessary for evaluating risk of producing polygenic traits like hip dysplasia.  You can effectively include vertical pedigree information without actually creating such a pedigree.

The farther back you go before you find a dog connected to a trait, the less that dog contributes to the risk.  Dogs that have a trait are the ones most likely to pass it on, parents less so and grandparents even less.  More distant ancestors are less likely with each subsequent generation to have contributed the necessary genes. 

How many generations to analyze behind your dog or a proposed cross depends on a number of factors.  If the trait is easily recognized (color, coat type) or commonly discussed (dentition) or if information is readily available (performance records), three generations might be enough.  However if recording of the trait is inconsistent or it is one people may not mention for fear of stigma, five generations may be more revealing.  Beyond five or six generations, likelihood that the trait could have been passed down without someone noting it becomes more and more improbable

When analyzing a pedigree, start with the first generation and move back along each line of descent.  Once you find an individual connected to the trait you are looking for, note whether it was affected, carrier or suspect.  Once you find something, do not proceed any farther along that particular line.  If you note more distant ancestors behind that one, you will inflate your result.   For example, if you find that the paternal grandsire has produced the color you are interested in, it does not matter that his sire was that color; the grandsire is the closest dog that you know had the gene, so skip to the paternal grand dam.  If she has no connection to the trait, go to her sire, and so on. Continue with each line of descent until you either find something or reach the last generation you are searching.

The next task is to determine the likelihood that you will get the trait.  These are not precise probability calculations.  Rather they are a method of consistent comparative ranking.  Who’s higher and who’s lower, with lower being preferable for undesirable traits like hereditary disease.  If you were looking at desirable traits, you would want high numbers.

I assign a value of 10 to an affected dog that appears as a parent on the pedigree, 5 to a carrier and 2.5 to suspect.  For each generation back, I divide that score in half.  

A suspect appears as a great grandparent –
2.5 (base score for a suspect) divided by 2 (grandparent) divided by 2 (great grandparent): 2.5/2/2 = .625 (which I round to 1)

Since gender isn’t an issue in this kind of search, values over 5 (equivalent to 50% in a percentage of ancestry calculation) can easily result.  It’s also possible to get values over 10 (e.g. if the sire was affected and the dam a daughter of a carrier, the score would be 12.5.)  Even though this procedure is similar to and based upon percentage of ancestry, the result is not a percentage calculation. I deliberately moved to a 10-point scale to help avoid confusion with percentages.

A pedigree I reviewed recently for epilepsy had a suspect parent (2.5), a carrier great grandparent (1.25), and five suspects on the 5th generation (5x .157).  This resulted in a score of 4.535.  Since the fractions of a point are not significant to the final result, I generally round to the nearest whole or, for anything under 1 up to 1.  For similar reasons, I have made 10 my maximum score—once you have reached 10 points you are extremely likely to get the trait.  If you prefer to use fractions, scores over 10, or a 100-point scale that is certainly acceptable so long as you are consistent in what you are doing.

 The Good Stuff

The above system works best for traits you don’t want and those that you do which are rare.  For good things that occur more often than  not you will want to determine the consistency of that trait across a horizontal pedigree or, better yet, in a vertical pedigree.

The modified percentage of ancestry method would work well for determining risk of hip dysplasia, but if what you want to determine is the likelihood you will get good or excellent hips in a cross, you would do better to look at hip scores across a vertical pedigree.  Having all excellent hips behind your dog is no guarantee he won’t produce HD.  But if there is little or no HD in the vertical pedigree going back two or three generations and the scores were largely good or excellent, odds are your dog will produce good or excellent hips in his offspring. 

Completing the Map

Follow the procedures outlined above for every trait you consider important.  You can use the results to determine potential weaknesses in a particular pedigree and map out a course for improvement.   A pedigree is very much your road map as a breeder.  Whether it helps you find your way or not will depend on how well you have taken note of the road signs and landmarks you’ve discovered along the way.