I "discovered" a chessie pedigree database today that contains COI values.

At what point does the Coefficient of Inbreeding become a relevant factor in evaluating prospective or actual breedings for chessies?

This inquiry may be the $64,000 question. I offer my thanks in advance to anyone who takes the time to enlighten me.

Hello Jefferson,
nice to hear from you.
What database did you read??
Did you read the articles I mentioned??
DD

Thanks for the response. I just noted the reference to the article(s) and will take a look. See the Chessie Database. Perhaps someone on this site could chime in about the database's history and reliability. I enjoyed playing around with the pedigrees. The Freezer x Douce breeding looks really nice.

COIs are calculated solely on paper, i.e., simply looking at dogs in the pedigree and calculating the COI based on how many times each dog appears on both sides of a pedigree.

Because the gene pool for this breed is small and has been since the Civil War (origial bottleneck), there is considerable inbreeding with any individual dog. In my private database, I am able to run COIs for 10 generation pedigrees, and have found values ranging from 6 to 56. There is no correlation between these values and quality of individuals.

Don't go by some arbitrary number; go by the quality of the individual dogs being bred together, as well as any production history of those individuals.

Lisa

If there are statistics amongst us,they good figure out in certain constant factors,let's say dogs,bitch and dogs, with fieldtrialtitles A,B,C.What was the COI of them and perhaps,you can predict that at a certain level of let's say p.e. 10 to 13 % of COI is the best.
Do not say that this is nonsense,look at the pedigrees in Chesapeake Roll Call.There must be a relationship,without any doubt.
DD

I believe, DD, that what you propose would only work statistically, if the COI distribution for the population were first determined.

It can't be done by the numbers.

When you look at successful FT breeding programs, the one common thread is selection. Each breeder knows what traits their line is strongest in, and selects prospects within the litter based upon their experience. Then getting those prospects into homes where they will maximize their potential is the next hurdle.

Another commonality among successful breeders is the ability to keep several dogs "on spec", and the intestinal fortitude to cut loose those that don't measure up. Selection works all along the process, from selecting the breeding pair, to deciding to pursue the AA career with the finished dog, or sell it as a gun dog.

Different lines produce well with different COIs. The Caroway/JJ dogs withstand tight inbreeding better than some. It is almost a closed line, as a Caroway-bred dog tends to produce better bred to another Caroway-bred dog than to an outcross.

Lisa

Lisa,
I'm speechless about your knowledge,that example about Caroway/JJdogs
Thanks very much!!!!
I have to study again.
DD

LVL --

What is "it" as in "It can't be done by the numbers"?

Statistically, what DD proposed initially is quite feasible if given COI values for a defined population of dogs. Do I read correctly that you performed a correlation analysis on your private database?

What value would this exercise hold for anyone? No one would know for sure unless the work was done and made available for people to make up their own minds.

Would I rely upon the resulting model in the selection of a puppy? Most likely not. Would the resulting model be of some influence? Perhaps.

My original question is when does COI become a relevant factor.

If your answer is that COI never becomes a relevant factor, I will accept that as your answer and believe your answer is incorrect.

If your answer is we simply do not have a enough data to know what COI values become relevant or that COI values themselves are meaningless, I could probably go along with either of those answers but would then ask why we calculate COI in the first place.

quote:

---

Don't go by some arbitrary number; go by the quality of the individual dogs being bred together, as well as any production history of those individuals.

---

- Please.

Jefferson writes:"If your answer is we simply do not have a enough data to know what COI values become relevant or that COI values themselves are meaningless, I could probably go along with either of those answers but would then ask why we calculate COI in the first place."

These are most likely calculated because they can. In a large population they might be meaningful. But consider that in the Chessie world there are only 1000 litters registered each year with a resultant 3700 pups registered. Also consider that the Chessie gene pool is so small and compact that every one of these 3700 pups can be traced back to a single common ancestor. And then try to relate that as to the intent of the COI.

Another consideration is that the COI can also reflect the reverse of what is implied here in this thread. It can also reflect the magnification of undesired traits that are passed along by these tight breedings. The whole trick in breeding is to try to offset or minimize the undesired traits while accentuating those desired traits you value. And in doing this there are always offsets, trade offs, and a large degree of luck involved.

T.Mac

quote:

---

It can also reflect the magnification of undesired traits that are passed along by these tight breedings.

---

Exactly. Thanks for your response.

As for what constitutes a "large" population for statistical purposes, I believe the population is large enough for meaningful analysis. The ability to trace to a common ancestor makes the COI parameter more, rather than less, significant a consideration. I disagree that the number is "arbitrary" as expressed elsewhere in this thread.

As it will state what I'm thinking about breeding,is that there a rational factors wich one can measure,and social wich one can see and emotinal wich one can feel.
It will always be a combination one can figure out more or less.
All people involved in this discussion are thinking about one or more things and that is quiet something for the goals we are striving for namely healthy,positive,good puppies,sound and a big contribution for the breed.
Big BUT,WE can think and discuss,but what about an ACC,who have time,money and people,in my opinion,to figure this out.
It's also for the ACC as representitive and responsible org. to work on this.Or am I wrong,plse tell me,and can somebody of you folks start a topic with teh ACC.
And one other thing ;hat means a small genetic pole?To what is this compered?Is 1000 dogs small???
DD

Yes, Jeff, I DID calculate COIs. Hundreds of them. This was done as an offshoot of something that Mario B. had discussed; his concerns that increased COIs might lead to decreased fertility.

Since there are so many things that can cause a bitch to miss, I did the opposite. Using AKC litter registration numbers, I "cherry-picked" litters with 12 or more registered individuals, input their pedigrees into my database, and then ran COIs on them all. I also ran some pedigrees of male dogs that I had first-hand knowledge of sterility as young dogs. There simply was no correlation between COI and fertility or lack of same. The largest litters (15+) came from a highly inbred group of dogs (COI .56), while some of the sterile animals had very LOW (COI .03) COI values. I think what is more significant is what INDIVIDUALS were these animals inbred on? The robustness of the animals being doubled-up on is more important than the raw score.

Lisa

Lisa -- Thanks for the clarification and description of your work on COI & Fertility. Do you know the history of the database mentioned in the post that started this thread? Are you aware of any other initiatives to determine whether a relationship exists between COI and desireable/undesireable characteristics? For example, has anyone looked at aggression as a possible function of COI? OR, is there concensus that agression is the result of a lack of socialization? Thanks again, Jeff

Lisa,
I agree with you about the lack of relationship of inbreeding and fertility. We learned in vet school that inbreeding might lower fertility, but my personal experience does not support this.

Mario

Precisely right, Mario. It is good to be able to put it to the test with real-life data. Being as the Chessie breed was bottlenecked not once, but several times in its development, we would expect there to be abundant fertility problems as a result of the intense inbreeding that was done immediately post the bottlenecks. The reverse is true.

I see all over the internet websites about how inbreeding will ruin purebred dogs, about how COIs are an indication of genetic diversity (or lack thereof). Nothing could be further from the truth. Many of those websites are created by animal rights or pseudo-AR folks. One of the most-quoted websites promotes the breeding of "Labradoodles". Draw your own conclusions!

One need only look at a Chessie FT or show specialty or supported entry show to realize that COI has no relationship to actual genetic diversity. No two look alike! While COIs in this breed can range pretty high, here is actually wide genetic diversity in alleles. Long may it remain so. The inbreeding is only on paper.

Lisa

Jeff;

Is that the ChessieData database?

Information was obtained by basically swiping Tom MacClanahan's info off his RollCall website (but they admit this in their introduction). There's a lot of missing info. COIs can change, depending on how many generations you utilize. I don't recall if they had a cutoff point, or if they just have COIs on everything because they can. Simply being able to generate a number is one thing. Being able to interpret that number in the real world is quite another. Interpretation takes experience, and these people who created this database just got their first Chessie last November. Not a lot of depth of experience there.

the history of my COI work was that Claire McCall contacted me with this question about COIs and fertility. As I recall, she and Mario had had a discussion during the FT in Maine, and it had got her wondeering. So, we decided to generate some numbers and put it to a real-life test. Claire knows I have a private database with full pedigrees on many dogs (some 30 generations long).

We first tested the actual accuracy of the COI function. Some pedigree programs have COI calculators that are highly inaccurate. Mine claimed to have been double-checked by a geneticist, but that wasn't good enough for me! Claire and I independently hand-calculated COIs for a selected set of pedigrees, and compared our values with what my pedigree program, and others produced. My program was, indeed accurate, so we went to step two.

Step two was to decide to what number of generations we would calculate COI. We settled on eight generations. It is too difficult to find full 10-generation pedigrees; while five or six generations don't always tell the whole story. We decided eight was the number.

Step three was scouring the stud books for sires and dams of litters containing 12 or more individuals. Not hard to do, given AKC's registration number scheme, and CKC's listing of litter sizes in their stud books! We discarded any dog that I couldn't generate a full eight-generation pedigree (no missing ancestors) on. I also discarded identical pedigrees. I had information on a handful of dogs that were documented with infertility of unknown cause. This is key, for there are many things that can cause infertility. I discarded any dog that was known to be infertile because of injury, disease, or hypothyroidism, etc., as these are conditions not controlled by inbreeding. When the dust settled, there were very, very few dogs that we did not know some underlying reason for their infertility. That's why we decided to look at the reverse of the question; those sires and dams of very large (very fertile) litters!

Your question about aggression is something that one simply can't nail down to genetics. There is so much involved with what makes up aggression. Not to mention, one person's "aggression" is another person's "good dog". All kinds of things contribute, including socialization (closely tied in with expectations, as in if a person wants a junkyard dog, he/she will do things consciously or subconsciously to make sure it happens!), then there are medical conditions that can cause temporary aggression. Yet, the novice owner with the dumbass vet (they do exist), may put the "aggressive" Chessie down without even performing rudimentary tests to rule out these conditions. A bad tooth or an ear infection can cause the mildest mannered dog to bite. Aggression? Or aggravating, non-stop pain? Hypothyroidism can also cause dogs to be grumpy, or downright aggressive. Lack of basic obedience training, of developing dominance over the pup can lead to problems. As you can see, none of this had Jack to do with breeding!

Well, I've rambled on enough. So now you know what scientists do to amuse themselves!

Lisa the number cruncher

LVL --

Thanks for the explanation of your work. I too have crunched a bunch of numbers in my career, but several years have lapsed since my last genetics course.

Are you also saying COI is calculated for chessies simply because it can be calculated? Tom M implied that was the case in his post above.

Are you saying COI values are meaningless for identifying certain predispositions in chessies? I named aggression just as an example. Either PRA or hypothyroidism would have better served as examples.

It sounds like your exercise with Claire McCall ruled out the possibility of "arbitrary" COI values.

My question is simply: At what point does the Coefficient of Inbreeding become a relevant factor in evaluating prospective or actual breedings for chessies?

"Relevance" could be on the basis of "health", field performance, . . . Do we even know if COI becomes relevant at any value other than a value of 1 (one)?

Jeff

Again, the COI is only an indicator, a tool. I don't think it is all that useful in THIS breed. Determining breeding pairs has far more to do with WHICH dogs the pairing will double up on, and whether those animals being doubled-up on are known to produce certain traits consistently. This has far more to do with knowing the animals and their production records, than whether a pedigree turns up a specific result on the COI.

yes, I believe COIs are calculated simply because we can. A more real-world number would be a modified Wright's, where you use known values of individuals in a pedigree to make predictions of each pup's likelihood of having a copy of that gene. I HAVE calculated such values for PRA. Here is where you are tracking a particular gene, rather than particular names on a pedigree. This is what I am trying to get across about COIs, they only look at names in a pedigree and spit out a number based on how many times ALL the names appear on both sides of a pedigree. This aggregate number isn't of much use. The modified Wright's COI is used to follow a gene. Thus, I will highlight all known affected PRA dogs in a pedigree. These dogs are a 1. All known carriers are a .5. Then trace it down from there. Since these can, and often ARE different dogs in the same pedigree, you will get a much more useable number than with the "bulk" COI. In other words, the COI is looking for reoccurrances of the same NAME in a pedigree. I am looking for reoccurrances of the same GENE, which can come from several different dogs named in that pedigree, and since we have known for many decades which dogs were carriers, they OFTEN are not doubled up on in a pedigree.

Wright's really is only good for tracking knon, single-gene autosomal recessives that have no interactions with other genes. Since so many of the traits we breed for (marking, memory, trainability) or against (hypothyroidism, hip dysplasia, aggression) are controlled by a number of genes that often have interactions with each other, the usefulness of the "bulk" Wright's COI becomes diminished.

I think the funniest application of COI was in a show-magazine ad for a top-winning Doberman. The whole slant of the ad was how this dog was a big winner, AND had a COI of 0. I don't care WHAT that dog's COI is, he still has MOST of his DNA in common with 99% of the rest of the Dobes. That is a breed that, no matter how much you outcross, there simply is NO genetic diversity. COI also is irrelevant in that breed. Except for advertising purposes, of course.

Lisa

LVL --

Thank you very much for the explanation.

I clearly see your point now and also see how COI could be and is misconstrued and assigned relevance when and where any relevance is buried in the complexity of multiple genes.

Thank you very much for taking the time to respond. You certainly seem to have a keen interest in this topic and have done quite a bit of related investigation.

I wish you well with your tests/trials this year.

Jeff