USING EBVs TO BREED BETTER DOGS
If somebody could tell you which of several potential sires was most likely to the improve a trait you're interested in selecting for, would that information be useful? Of course, this is what you're trying to figure out in a subjective way when you dig through pedigrees, and if there's more than one trait to worry about it can get pretty complicated. There is a better way to do this.
Let's say you're worrying about good hips. Dogs that have good hips, and that have lots of relatives with good hips, can be said to have good "breeding value" for that trait. If better hips is what you're looking for, the dog with the best breeding value is most likely to have the genes that will produce good hips in offspring. More often, a particular dog might have good hips but some relatives will have better hips than others, and estimating a particular dog's breeding value gets trickier. How can you make a reasonable prediction of what a particular dog is likely to produce if there are dogs with both good and bad breeding value in the pedigree?
A statistical technique called "estimated breeding values" (EBVs) allows you to attach an estimate of breeding value to all of the dogs of interest that reflects not only each dog's phenotype but also its genotype. Any time you get more information about the relatives of a dog of interest - it has a litter of puppies, or a sibling has a litter, or one parent produces a litter of half-sibs - you can incorporate this information into your estimation of breeding value and improve the accuracy of the EBV. As you accumulate more and more information about the dogs in the pedigree and their relatives, you are able to make better and better predictions about what any particular dog will produce.
There is an excellent introduction to using EBVs in dog breeding (Cavalier King Charles Spaniel) here, and more information here.
You can see EBV data for hip and elbow dysplasia that was derived from the OFA database here.
EBVs have been used for decades to guide selective breeding of guide dogs, here.
ICB is in the process of implementing program that will be able to compute EBVs using OFA data or any other database with sufficient information. We will be building this page as a resource to support this and to provide information to breeders on how this information can be used. Watch this space!
Let's say you're worrying about good hips. Dogs that have good hips, and that have lots of relatives with good hips, can be said to have good "breeding value" for that trait. If better hips is what you're looking for, the dog with the best breeding value is most likely to have the genes that will produce good hips in offspring. More often, a particular dog might have good hips but some relatives will have better hips than others, and estimating a particular dog's breeding value gets trickier. How can you make a reasonable prediction of what a particular dog is likely to produce if there are dogs with both good and bad breeding value in the pedigree?
A statistical technique called "estimated breeding values" (EBVs) allows you to attach an estimate of breeding value to all of the dogs of interest that reflects not only each dog's phenotype but also its genotype. Any time you get more information about the relatives of a dog of interest - it has a litter of puppies, or a sibling has a litter, or one parent produces a litter of half-sibs - you can incorporate this information into your estimation of breeding value and improve the accuracy of the EBV. As you accumulate more and more information about the dogs in the pedigree and their relatives, you are able to make better and better predictions about what any particular dog will produce.
There is an excellent introduction to using EBVs in dog breeding (Cavalier King Charles Spaniel) here, and more information here.
You can see EBV data for hip and elbow dysplasia that was derived from the OFA database here.
EBVs have been used for decades to guide selective breeding of guide dogs, here.
ICB is in the process of implementing program that will be able to compute EBVs using OFA data or any other database with sufficient information. We will be building this page as a resource to support this and to provide information to breeders on how this information can be used. Watch this space!
