[PyrNet-L] breeding methods

[Kshoffman@aol.com]

In a message dated 12/20/99 1:11:30 PM Eastern Standard Time, 
Dancingpyr@aol.com writes:

<< Linebreeding, from my limited knowledge of the subject, can be
 very useful in establishing and maintaining desired characteristics.
  I think where it goes wrong is when uneducated "breeders" mistake
 inbreeding for linebreeding.  Then it can go horribly wrong.>>

Linebreeding is more or less just a slower form of inbreeding.  Both cause 
the same end results  -- both cumulatively increase the level of homozygosity 
(like pairs of alleles), both cumulatively decrease the amount or variation 
in genetic material one has to draw from, both usually are an attempt to 
capture a concentration of genes from one or more admired ancestors.

The difference between the two methods is the rapidity with which the method 
fixes like pairs of some alleles while discarding other alleles altogether -- 
the latter applying to both intentional selection against major deleterious 
mutations and visible undesirable cosmetic traits, and the concurrent 
inadvertent discarding of unrelated genetic material via the loss of whatever 
alleles happen to lie on the same chromosome as those being purposefully 
selected against.

The dog genome is estimated to include 80,000-100,000 gene pairs; 
160,000-200,000 alleles. These are arranged in a linked fashion on the 78 
pairs of chromosomes. This amounts to an average of ~1100 alleles per 
chromosome. One set of chromosomes may include gene pairs that code for 
hundreds of different unrelated traits, processes, pathways, and functions. 
Multiple genes that work collectively and/or interactively to determine the 
expression of numerous traits, processes, pathways, and functions may be 
spread across several different chromosomes, one piece of the equation here, 
one piece there. Some alleles on a given chromosome may be beneficial 
alleles, some may be deleterious alleles, yet all aligned on the same 
chromosome in a linked fashion and inherited as a group, one set of alleles 
from each parent.

Every individual carries from a few to several to dozens or more hidden 
deleterious alleles that impact on health and fitness in one way shape or 
form. These may be linked on the same chromosome to other very desirable 
visible traits. Some of these may be major alleles that can have a 
deterimental effect on their own if combined in double dose, some may be 
minor alleles that independently don't have much effect, but collectively, 
when concentrated in vast numbers towards either a plus or minus direction 
(for example, optimal versus sub optimal) can have a tremendous positive or 
negative impact.  Inadvertently concentrating vast numbers of minor 
suboptimal alleles via continuously limiting the gene pool one is drawing 
from will eventually cross a threshold which results in expression of 
impairment of function to varying degrees.

Regardless of how talented a breeder is, or how good a researcher a breeder 
is, they can't possibly know about every single allele or combinations of 
alleles an individual dog codes for that may contribute to the expression of 
genetic defects and diseases and varying grades of impairment of function.  
If it were so easy to work up such a code, then the whole process of mapping 
genomes would already be completed and we would be well on our way as 
breeders to "mixing and matching" accordingly.  Unfortunately, we are far 
from there yet, because it just isn't quite so simple.

Another factor to consider: It is now estimated between .5 and 1.0 *new* 
deleterious mutations occur per each individual per generation.  So, no 
matter how hard a breeder tries to "bring defects to light" via linebreeding 
or inbreeding, it is never going to be possible to purge *all* the 
deleterious alleles while still maintaining enough beneficial genetic 
material to maintain long-term viability.  In essence, eugenics is the 
impossible dream.

Dogs that are more closely genetically related, i.e. from the same line or 
family, are much more likely to carry the same deleterious alleles. These 
will eventually turn up in a linebreeding program just the same as they do 
with an inbreeding program, it just takes longer (more generations) to 
realize this result via linebreeding. What might be destined to happen in 10 
generations of full-sib breedings (possibly even extinction of a line) may 
take twice as many generations or more of matings of a slightly further 
removed genetic relationship.  For these reasons, linebreeding is generally 
considered to be the "safer" method, because it slows things down.

As compared to close inbreeding, linebreeding reduces the risks of major 
deleterious mutations combining in double dose, delays the onset of 
inbreeding depression, and allows slightly more variation in available 
genetic material to draw from.  However, in comparison to linebreeding, 
outcrossing is even safer for all the same reasons because it puts even more 
genetic distance between two mates and thus slows things down even further.  
As compared to linebreeding, outcrossing even further reduces the risks of 
major deleterious alleles combining in double dose, avoids the onset of 
inbreeding depression, and it also reduces the risk of vast numbers of minor 
suboptimal alleles being concentrated as they so easily can be 
(inadvertently) via familial type breeding methods.

The terms inbreeding, linebreeding, and outcrossing as generally used by dog 
breeders are rather arbitrary in a sense.  All purebred dogs in the big 
picture are by necessity inbred to some extent.  This is because entire 
breeds are descended from a small group of "founders" and to that end, all 
members of a given breed are related if one takes their pedigrees back far 
enough. One breeder's inbreeding might be another's linebreeding.  One 
breeder's linebreeding might be another's outcross. There really aren't clear 
and concise definitive definitions to differentiate one from another. Some 
full sib matings may result in a lower level of inbreeding than some cousin 
matings. This is because inbreeding is cumulative over generations.  So if a 
cousin mating has several other familial breedings behind it, but the full 
sibs' parents were more or less unrelated, then the offspring of the full 
sibs are going to be less inbred, less genetically related, than the 
offspring of the two cousins.

The more accurate gauge of level of inbreeding is Wright's Coefficient of 
Inbreeding formula.  Many pedigree programs allow this computation, but it 
can also be done manually.  Wright's coefficient of inbreeding measures the 
estimated level of homozygosity (like pairs of alleles). This computation 
results in a number between 0 and 1, usually reported as a percentage.  The 
closer the computation is to "1", the more inbred, the higher the level of 
homozygosity.  The closer the computation is to "0" the less inbred, the 
lower the level of homozygosity.  It's all a big continuum. So called 
"linebreedings" might yield coefficients of inbreeding ranging from as low as 
perhaps 5% to as high as perhaps 60%.  It all depends on one's interpretation 
of the terms.

There is a mountain of peer-reviewed scientific evidence to suggest there is 
a statistically significant correlation between inbreeding coefficients, 
defect rates, and fitness traits, i.e. immune system function, metabolic 
function, reproductive function.  As inbreeding coefficients rise, so 
generally does the frequency of defects and loss of or impairment of fitness 
traits.

Kelley Hoffman
kshoffman@aol.com