Phenotype or Genotype when breeding?

[Mark Billadeau]

Pam,

One needs to be very careful about how "the best" is defined so that we are not increasing the level of in-breeding in our gene pool. If the defintion of "the best" reduces the number of dogs worthy of being breed to too small a number we'll be in trouble.

Mark

[Liz P]

I am sure that Pam is commenting on the fact that too many people are breeding dogs that aren't very good on stock (can't even make a living with a set of predictable farm chores, or worse, yet, are completely lacking in instinct).

[Mark Billadeau]

I'm sure Pam was thinking along those lines, but others do read this.

[Liz P]

It's certainly a balancing act; maintaining high standards while saving as much genetic diversity as possible.

[Gloria Atwater]

Thanks for the interesting discussion, everyone!

And thanks especially to Zach for explaining genes in terms that even I, who doesn't know genotype from granola, can understand. Thus, as someone who is not a breeder but who is a buyer, I'll toss my two bits in the mix.

 

Given the hypothetical scenario below, and presuming your bitch was one I really liked, while also presuming both sires exhibited qualities that would equally compliment the bitch ... then I would be most tempted to look at the average worker who has produced outstanding pups. If he's been put to various bitches and his pups are consistently good workers and of a consistent quality that appealed to me, I'd be thinking seriously about pups by him. He's proven and furthermore, his grandparents are also probably knowable and I could find out things about his lineage, what sort of dogs they may have been and whether that line would suit me.

(Besides, what's "average?" Maybe he's never made the double lift at Meeker, but if he's doing his job in a competent, workmanlike and intelligent manner, and not pulling sheep down or running them through fences, average can be pretty good.)

The second dog may be awfully spiffy on the trial field, but who is he? What's behind him? What were his grandparents like? What are his siblings like? Is he an outstanding trial dog because he's got that natural brilliance, or because he's been handled well? Like Secretariat, who's been invoked here, he may or may not be able to produce himself and there's nothing behind him to tell me whether his line are natural out-runners or prone to dirty grips or whatever. All I see is this one dog and whatever he expresses all by himself.

Therefore, as someone who might be looking for my next trial/working dog, I'd be most tempted by the sire whose pups are of known quality and have been consistently so from several different dams.

There. My totally unscientific opinion, for what it's worth.
Best regards,

Gloria

 

 

Jovi and others,

 

Excellent points. But lets say that both potential sires are similar in size, conformation, health and temperament. Both having equal complimentary characteristics to my bitch.

 

One is an average worker but has produced many outstanding trials dogs with different dams.

 

One is an outstanding trial dog but has a pedigree of unknown and unproven dogs and has never been bred.

 

I love how you mentioned probability. If only there were a software program for that? :~)

 

I guess it's our in our heads and our gut instinct.

 

Bonnie

[wyndrunhr]

Gosh, this is a GREAT discussion, from many experts from differing backgrounds!

 

I think I get it Zach and have heard this before, BUT

 

If the Stud dog with the less than perfect phenotype has a history of producing supreme phenotype in his puppies from different dams, is that not just as strong or stronger indication of the stud's genotype than his actual phenotype?

 

Bonnie

 

I'm not a breeder, but I teach genetics as a small section in my biological anthropology class. I think the key that some here might be circling is this....

 

First, you can't know the gene you're after with today's technology, you'd need some serious (read: federal) investment to even begin to look at the actual genes (probably several interacting sequences of genes) that have anything to do with the behavior you're interested in. So you can almost forget about the genotype. The phenotype is the only solid indication that the desired genotype is in the dog. And, since the desired phenotype might be the outcome of several possible sequences (different combinations leading to the same effect), you might have two perfect dogs with different gene sequences and the same phenotype.

 

Second, genotype can change from one generation to the next....in fact it has to (recombination)....and sometimes you get a new sequence in the new generation (f2) that has the result of suppressing the phenotype in the previous generation. It might not be gone, but it isn't coming out in the new generation. The subsequent generation (f3), however, might lack the suppressing element int he code and suddenly you get the phenotype again.

 

SO....Without knowing the genes though, and we don't, no one does, you can't be sure if the desired sequence is in that second gen (f2) dog. the third gen (f3), with the desired phenotype, might have it because 1) it was suppressed in f2 and not it f3, 2) the desired sequence came from the other parent, 3) breeding the f2 dog and its mate spontaneously generated a different sequence with the same phenotype as was seen in f1. This means, the only way you're sure you have a dog with the desired genotype, is that it exhibits the desired phenotype.

[Pearse]

Of course we can't truly know genotype, but the results of previous matings can give you some clues as to genotype based on the phenotypes that the dog produces (see windrunhr's comments).

 

When choosing a stud, I like to know that the dog has produced quality pups. That is key. The best dog in the world might not reproduce well.

 

 

No. We can absolutely know genotype with 100% certainty. We can sequence the genome.

 

What we can't do, is look at the genotype and tell anything useful about what kind of a stockdog that dog will be.

 

If we sequence enough genomes of enough dogs whose working ability is known, we would eventually be able to figure out which portions of the genotype were involved in working ability.

 

Would that enable us to predict working ability in the offspring of two dogs whose genotype was known?

 

It depends.

 

It depends on how many genes are involved in producing working ability, how many alleles (variants) of each of those genes are present in the working dog population, how close those genes are to one another in the genome (the further apart, the greater the recombination rate), and whether their effects were completely expressed (100% penetrant) or diluted by the effects of other genes/alleles.

 

In addition, we now know that certain regions of the genome can be activated or shut down by non-genetic (epigenetic) means, and that those changes can be heritable, so even if you have the genotype, you may not be able to predict phenotype based on that alone.

 

And, even if you knew all of that, you'd still need to sequence the genome of all offspring.

 

And, even if you knew all of that, you'd still need to make sure that the dog was exposed to the proper training at the proper time, raised in the correct environment, and correctly fed and exercised.

 

So, for the forseeable future, the only way to pick a stockdog is the same way we've been doing it for centuries. Breed the good 'uns to good 'uns and hope for the best.

 

Pearse

[Liz P]

What I should say is, your average person can't know the genotype right now because it's too expensive and wouldn't give them any valid information. In the future it might be possible, but I was talking about right NOW.

 

No. We can absolutely know genotype with 100% certainty. We can sequence the genome.

[Diana A]

The whole point of knowing the genotype (assuming someday we have the technology to do it and the knowledge to know what we're looking at) is to predict what a particular pairing of dogs will produce. If the stud has a track record of consistently producing a certain quality of pups when used with different bitches, then that to me would be the important thing if was planning to breed and making decisions about what stud to use. With any breeding you're trying to predict - if I put these two dogs together, what will I get in the pups? Will it be what I want? You can start with phenotype - wondering if the stud will produce what he is. Even better is after he's produced some pups, see what he actually produces (it may not match his phenotype). To me a stud who consistently produces traits I like even when put to different bitches with different traits would absolutely win out over a stud who was himself very nice but didn't consistently produce what I wanted. Somewhere way down the road when we've isolated genes for eye, balance, power, etc you may be able to sequence sire and dam and make a good prediction of what you may see in the pups, but given that all of those traits are likely controlled by multiple genes, that day is a long way off. I got my last pup from a third-time repeat breeding for just this reason - I had a pretty good idea exactly what I was getting based on the previous two litters - no trying to guess how they might turn out. I didn't need to know the genotype, and didn't even know the phenotype of the dam that well, but I did know what that pairing was producing and that was the piece I cared about. So if you already know a stud is producing well, trying to figure out actual genotype is secondary, in my opinion.

 

Edited to add: I am not a breeder but I consider these things when looking at a potential breeding to get a pup from. Since you can't tell from a pup what they will be as an adult, the best you can do is look at the parents and try to guess what you'll be getting.

[Pam Wolf]

Yes, Mark and Liz that is what I was talking about. Far too often people are breeding for $$$ or the fact they "love' their dog and want pups from her/him, or they have had some success in low level or mediocre trials and breed them (usually for the above reasons). Or they breed them because they can produce (or to) a particular colour.

 

Running in Open does not always mean a dog is breeding worthy IMO. Nor does the fact the dog works on a farm/ranch. If you cannot see quality then you should not breed or seek the help from someone experienced with knowledge and ability to see quality before breeding.

 

And Mark, yes, we have to rely more on phenotype, but also the heritage of the dogs (the ancestors). And external factors CAN alter the potential of a dog. Poor training, stressful environments etc can affect the working phenotype of a dog. It can be difficult to see what the dog may have been or is capable of in many cases.

[gonetotervs]

Pam and others, I 'm not sure I agree 100 percent. The number of dogs that achieve high titles, have the structure, health and overall temperament to be outstanding representatives of their breed is what -- the top 5 percent? Top 2 percent? I would dearly like to buy my puppies from this select group any day of the week.

But if we only breed those, then our inbreeding will eventually produce a very limited gene pool with less desirable consequences. Every really spectacular working dog that has no "pedigree" probably has a breeder who thought that their not so great dog was worth breeding -- and I can accept that a bit better than breeding for bucks. But genetic diversity is not always achieved by breeding just the best to the best.

And I'm not a breeder and don't know how to square that circle in real life.

[Liz P]

I've heard genetics experts advise breeding 1/3 of dogs to maintain selection pressure while keeping genetic diversity. I don't know the math behind this recommendation.

 

The problem lies in the large number of pet/show/sport bred dogs that lack enough instinct to make a useful working dog. The breed is defined by the work, so I don't count those dogs as part of the breeding population. That leaves us with farm and trial dogs. There is nothing wrong with breeding the farm dog that never trials as long as it is a truly useful working dog (not talking about moving animals through paddocks of hobby farms). (What % of ABCA Border Collies actually work?)

 

From a genetics standpoint, I don't agree with ONLY breeding the dogs that qualify for the National Finals. If we do that, we are in serious trouble. However, if we don't have work as a standard, we will lose the usefulness of the breed. The only real solution I can think of for choosing a breeding is to either know what to look for when watching the parents work or have a mentor who can help you.

[Tea]

Yep, I agree with Julie and Mark and Pam.....Gloria- I think Sweep has Granola instead of genes.

 

I will add, tho it may have already been stated- I think it is important-

 

What kind of trials? What level?

Hobby farm flock of 100, or Soldier Hollow?

 

What Kind of work, what level?

Moving 1000 head of range sheep, Loading 200 head of goats?

 

Separating calves from sheep?

 

Moving mama cows? Breaking range cattle without loss of fences or tempers?

 

 

 

 

What kind of stock?

 

Katahdins? Targhees? Marinos? Scotties?

 

Angus or baby jerseys? Corrientes? Highlanders?

 

Ewes and lambs? Mama cows with day old calves

Bulls? Flock of Ramboullet rams

[Denise Wall]

Not to hijack this topic but since this keeps coming up:

 

Pam and others, I'm not sure I agree 100 percent. The number of dogs that achieve high titles, have the structure, health and overall temperament to be outstanding representatives of their breed is what -- the top 5 percent? Top 2 percent? I would dearly like to buy my puppies from this select group any day of the week.
But if we only breed those, then our inbreeding will eventually produce a very limited gene pool with less desirable consequences. Every really spectacular working dog that has no "pedigree" probably has a breeder who thought that their not so great dog was worth breeding -- and I can accept that a bit better than breeding for bucks. But genetic diversity is not always achieved by breeding just the best to the best.
And I'm not a breeder and don't know how to square that circle in real life.

 

Although I originally wrote this dart board analogy to describe what I *think* might happen when working breeds are lost, it applies to trying to imagine how to keep the working gene pool healthy. I am a breeder (or was at least) and it's a very difficult to circle that square in real life. Many here will be familiar with this analogy already:

 

Assume the border collie is the theoretical breed, where many strong workers existed in the original breeding pool and the need for their work was not lost or reduced over time but instead the dogs became less and less useful for it.

I believe it simplifies the concept I’m trying to get across to think of the different levels of workers in concrete groups, even though, in reality, the scale from all to none is on a continuum. And, in reality, each dog of a breeding pair should be evaluated through actual stockwork for each of the many traits involved, and bred to the best complimentary mate in an effort to produce the proper mix of these traits in the progeny. So, this analogy is strictly my theoretical attempt at a simple representation.

**************************************************************

Imagine something such as a dart board, with a bull’s-eye and several circles that indicate areas farther and farther from the middle target. Let’s say the bull’s-eye circle is red, the next circle is orange, the next yellow, and the very outside circle is white. The actual area within these circles varies depending on the number of dogs in each class at any one time.

Now let’s define the groups of dogs within the different colored circles. Please remember all of these categories in this hypothetical situation represent the genetic potential of these dogs. In other words, this is what's in the gene pool. I'm not talking about what people think the dogs are or don't know whether they are or not due to not having tested them:

Red circle (bull’s eye) = The very best quality of working border collies. A working definition might be dogs that are exceptional enough to save a great deal of time and manpower for a livestock operation.

Orange circle = Useful dogs who save time and manpower for the operation but who are not top quality.

Yellow circle = Dogs who will work a little, but wouldn’t be considered useful workers on a real livestock operation because they would cost time and cause too much trouble to train or use. IOW, someone may want to pretend they're actually helping, but they really aren't and sometimes they're hindering. Although they may show some herding instincts, it's not the right total package for work.

White circle dogs = Not interested or not capable of doing anything with stock except maybe chasing or showing only prey drive. So, not useful or way less than helpful, and sometimes downright dangerous to the stock.


Livestock working ability is comprised of many complex traits. These traits all need to fit together just right and in the right amounts for the dog to be the complete package, and be considered a top worker -- the bull’s-eye. Achieving this package with the consistency needed requires stringent evaluation and selection for working ability every generation. Because of the complexity of reproducing behavioral traits such as these, it’s difficult to get this package that is a top worker, in every pup, or even close, despite crossing the best to the best. This is partly because some dogs, for whatever reason, aren’t good breeders, no matter how good they, themselves, are. So let’s say if only red circle dogs were crossed, only 80% of that number of red circle dogs would be produced in the next generation. (This is a hypothetical number – it may actually be more or less.) Therefore, breeding only red circle dogs will not replace all of the red circle dogs, and the number of red circle dogs will drop each generation if only these crosses are used.

As with other breeds used for other purposes, many a top sire gets bred to a mediocre bitch. Because the working genes were (are?) still highly concentrated in the border collie gene pool, the chances of hitting upon a dog that may not be a top worker herself but is a good breeder, are still pretty good. This type of good breeder would be mostly in the orange circle with a few in the yellow circle, but almost none in the white circle. Breeders of these top working sires may take a stud pup from these crosses to increase their chances of hitting on a good breeder should their top bitches not be, or not cross well with their choice of stud dog. In other words, the top breeders still rely on the peripheral pools of dogs that are not as good workers themselves but are good breeders, to provide some of their next generations of top red circle dogs. As long as the emphasis is on breeding for work and the momentum of most of the breeding is going toward breeding for the bull’s-eye and concentrating only the working genes, the number of red circle dogs will be replaced each generation and maybe even expanded.

Now, suppose the breed becomes popular for dog shows, pets, and dog sports such as agility. Suppose these people do not only buy puppies from the working bred dogs. Now instead of a mostly dead end gene pool -- dogs that will not be bred but only used for dog sports, etc., these dogs with no working ability will be bred as the demand increases. The number of white circle dogs increases. And since people seem to want to claim their “borders” can still herd with the best of them, or the sport dog people need to tap into the working traits for success in their endeavor, they will look to the working circles for breeding to try to get these traits in the pups. Regardless of how it happens, however, now the momentum has changed and the working genes are being diluted, instead of concentrated, in this peripheral gene pool that has formerly been the source of good breeders to help replenish the red circle top workers. As this trend progresses, the good breeders in the peripheral gene pool become rarer, the yellow circle fades more to white, the orange fades more to yellow and the red fades more to orange. Unable to replace themselves without the help of the strong working genes formerly present in the peripheral gene pool, over time, the number of dogs truly in the red circle diminishes until the gene pool is too small.

[Tea]

Always a great read, thanks!

[Pam Wolf]

The 'best' leaves a lot of room for interpretation when speaking of working dogs. Is it the top trial winner, maybe; is it the backyard pet who doesn't even look at a moving object- most likely not. The best IMO is a dog quite capable of handling livestock under various conditions requiring instinct (over) and commands (showing trainability). There is room within this for a wide variety of dogs, those that are quite biddable and those that are very natural.

 

What is not needed is breeding those dogs which may be wonderful pets, but have nothing to offer the working dog. We do not need more 'fluffy' dogs either (and even fewer of the 'hybrid show/sport/working' types). We do not need dogs selected for SAR, tracking, drug detection, hearing ear dogs etc being bred and CALLED Border Collies. While these later are worthwhile pursuits, it is NOT a test of a dog's working abilities. Nor is prancing abound a ring staring at liver, nor is competition in small arena trials on school sheep. Any good working dog could be trained to do all these pursuits-well, maybe not staring at liver, but certainly staring at goats .

If you want to breed the wonderful dogs called "Border Collies" then make an effort to learn what they are, not some pseudo version of them. Make an effort to have every dog you breed 'work tested'.

[Sue R]

Yes, thanks, Denise! And thank you, Pam.