Was having a nice chat with another Team member this weekend, and the topic of pedigree analysis relative to individual dog performance came up. Specifically, we were talking about polygenic traits, which are so hard to predict.

While the pedigree is a fine tool to help guide a breeder/owner, it is just a tool, and should not become the driver! We talked about his dog, within whose pedigree lurks a couple of individuals who either barely passed OFA, or did not pass at all. The presence of these dogs in the pedigree would indicate using the dog cautiously *at first*, until his own production abilities are known. For instance, if I had a bitch that I knew produced dysplasia, I might not breed her to him until I knew what his "batting average" regarding hips was.

Well, I ran some "quick-and-dirty" numbers on this dog's production. Of the 100+ offspring with OFA numbers, 36% have passed with excellents! The average for the breed is 10% excellents, so this dog, despite the pedigree, is producing excellent dogs at three times the rate of the average. Now if I had a bitch with poor hips behind her, or knew that she had produced a dysplastic pup, would I use this dog? Better believe it!

This is an excellent example of knowing the limitations of breeding systems. Some people are slaves to the pedigree, believing they can predict a dog's potential for producing this-or-that genetic problem just by studying the pedigree. Some people even go so far as to tell folks who they "should" or "should not" breed to, based solely on pedigree. This is wrong! Especially with polygenic traits like hip dysplasia, elbow dysplasia, etc., it's just too hard to predict.

Say for simplicity's sake, there are ten different genes controlling hip conformation. If a dog has five good genes and five bad genes, he or she may just barely pass OFA. Yet, he or she only passes on half of their gene set on to the next generation. A small percentage of the sperm (or eggs) that such a dog produces will contain ALL good hip genes, and NO bad hip genes. Statistically speaking, HALF of that dog's sperm (or eggs) will carry MORE good genes than bad, while half will carry more bad genes. It is a crap shoot as to which one will go into making any particular puppy. So, just because there is a dog in a pedigree (especially several generations back) with questionable hips, should you refuse to breed to the descendant? We all make choices, and have to keep the whole animal in mind when choosing breeding pairs. Sometimes you just have to gamble, and sometimes the gamble works out!

The other thing I don't like about the "pedigree system" of breeding dogs, is that all too often, it is a system that focuses on avoiding negatives. In an effort to keep from producing dogs with ANY kind of problem, many times the goal of producing superior specimens gets lost. Instead of breeding TOWARD an improved specimen, one breeds AWAY from a negative trait. So a level of mediocrity often results. Sure, the dogs look good *on paper*, with many generations of OFA, CERF, etc. , but the individuals that such paper represents are often middle-of-the-road. I think it's far better to select breeding pairs based on how well those two individuals suit each other. BOTH sides need to bring something to the table, as individuals, not just on paper. You can't rely on the genes being there just because the pedigree looks good.

Lisa

Lisa,
I can only partially agree.
Your example of the a dog with 100+ offspring works well for males but the dam will usually not have numbers of that magnitude to allow looking at the individual dog's track record.
All pedigrees and puppies and gambles. Add in the polygenetics, variable expression, systemic and enviromental factors why not stack the deck (pardon the pun) with known quantities?
Selecting a breeding by what both bring to the table is essential but unless it is a repeat breeding all we have is certifications and titles.
Tim

Yes, bitches are at a distinct disadvantage. But then, so are males who have a "black mark" somewhere deep in their pedigree. The "pedigree snobs" will point out this "black mark", often to the uninitiated, and often to their own advantage ("Don't breed to that dog, he's got XYZ in his background, breed to my stud instead.") In order to get past this "black mark", the dog has to be so outstanding in his own right, that people will use him *anyway*, which was the case with the dog I mentioned. I wonder, had this dog not attained the titles he did, would he have been bred to? Or as much? Likely not, and simply because of something 4-5 generations back. This is where "pedigree snobs" can do damage in a breed with a gene pool as small as ours, eliminating a dog from the gene pool because of *one* thing far back in the pedigree.

I think it would be very helpful to the Chesapeake community, considering our very small gene pool, if we had more access to pedigree *breadth* information, rather than just pedigree depth. If a dog on a pedigree was only OFA fair, yet his littermates were excellent and good, he is a better individual behind a dog than one that is excellent from a litter of otherwise dysplastic pups. Likewise, if that OFA fair dog was known to produce a proportion of goods and excellents, perhaps above the breed average, then the appearance of that dog in a pedigree should not give one pause.

Now, I know that GDC and OFA have combined forces. I also know that they are going to revamp the OFA database. Let's hope that they include a "reverse search" feature, where one can type in a dog or bitch and find out what that animal has produced. One could also find out info on littermates, etc. This is far more useful information for someone trying to fit those pieces of the puzzle together. But this is only one trait! Too bad there is no way to track other intangible, polygenic traits, like marking, trainability, drive & desire, and the like! Now, THAT would be a database worth having!

Lisa

Hey, Lisa,

There is a reverse-look-up facility in the current OFA website. Down toward the bottom of the page, you can get information on the clearances that the animal produced, as well as information on its sire and dam. Unfortunately, I have found that information on sire and dam AKC numbers is the least reliable (and often unavailable on OFA) part of the database.

For relatively recently born dogs, I search on what I'm going to call the "basic number" for a litter - everything in front of the slash. Then you can find out all of the passing clearances for that litter.

Happy hunting, Claire

The monkey wrench is the old AKC #, and dogs that come from other countries, that have their OWN litter #'s from AKC.

I just had fun checking CERF's site, they now have the CATEGORIES (if any) posted. It's interesting.

PB

Yeah, and some people forget to bring their dog's AKC slip with them, so the OFA number is attached to a dog with a name like "Chester", no AKC #, no sire & dam info...

It is interesting about the CERF categories, though. I think these refer only to those categories that persist, though. For instance, at one time, Gypsy was diagnosed with those pesky punctate cataracts. These have never re-materialized through many follow-up re-CERFs, so I can only surmise that CERF removed her Category "E" rating, although she still has her Category "A" rating (she has distichia). It's interesting to see how many dogs DO have those small cataracts. Seems pretty prevalent in the breed.

Lisa

I think people are getting more educated about how to do these kinds of things.
I.E. when I took Jack in for his ofa x-rays three years ago, I didn't know about having to have permanent ID for it to be on AKC documents. --Or even to get his elbows checked while he was out.

I really like seeing the rise in OFA #'s. For example almost 3 years ago Jack got CB-6288, my last x-ray just this past August was 7149. Peake's are getting hip numbers at the rate of about 1 dog per day. It seems to be catching on, but it is still only about 6% of the rate they are being born.
Elbow #'s are climbing as well. It also seems to me that people are more willing to release all information good or bad.

quote:

---

Too bad there is no way to track other intangible, polygenic traits, like marking, trainability, drive & desire, and the like! Now, THAT would be a database worth having!

---

Preach on Sister Van Loo!

PS Gypsy only shows Category A. 10 CERFs? That has got to be a record...

[This message was edited by Paul Brown on December 11, 2002 at 06:35 PM.]