MANAGING   "INBREEDING"

UNTIL YOU IDENTIFY THE REAL DISEASE, YOU MAY NOT BE SUCCESSFUL AT A CURE

At the beginning of the AI industry, when most dairymen used natural service, the selling points for the new technology of artificial (manual) insemination were:

(1)   Reduce or eliminate the risk of injury or death by bulls;   (2)  Have more control over infectious disease, eliminating veneral-based diseases;   (3)  Have some choice in genetics, avoid having all your replacements sired by one bull of unknown merit;   (4)  Put a productive cow in the bull's stall and add to dairy cashflow;   (5)   Obtain more precise breeding records, so as to better anticipate when to dry off cows and when to expect calvings.

Implicit in the above reasons-- related to genetics-- AI gradually opened up access to seed stock from around the world.     You were no longer limited to the breed and bloodline preferences of breeders in your locale, you could cross bloodlines (and breeds) at will-- especially when, after twenty years of "fresh" semen, the option of frozen semen (first sold in the 1960s) increased sire access exponentially, and inaugurated a higher level of genetic competition.

Linebreeding-- the major selection tool prior to indexing

"Linebreeding" and "Inbreeding" were popular among most purebred breeders in the 1930s, 1940s and into the 1950s, the period when AI stud services were organizing and expanding.   These mating approaches were based on the desire to replicate the better ancestors in a more consistent way, and evolved from the experiences of the more successful stockmen of the nineteenth century (prior to milk testing or type classification, let alone AI).  

With linebreeding, you mate related animals so as to maintain the percentage of descent to some individual superior animal at a constant level.   With inbreeding, you create much closer levels of relationship-- brother to sister, sire to daughter, mother to son-- ie, you limit ancestry within a family.     

In the case of linebreeding, you achieve some homozygous pairing of genes, which influences consistency in the trait focuson which the selection of ancestor was based; but you also could maintain a mostly heterozygous pairing of genes, due to having a multiple of ancestors in total (ie, sharing one grandsire or grandam within a mating still allows the other grandparents to be different, thus providing potentially different-- heterozygous-- genes into the conception).

[continued next page] 

Inbreeding-- can be carried too far

It is the consensus of any number of researches, beginning with an amazingly close breeding experiment done by USDA at the Beltsville Experiment station in the 1940s, involving many different breeds, repetitively mating sires to daughters, that inbreeding after some generations usually will result in a loss of "vigor"-- ie, you will at random begin to experience lower fertility, more stillborn calves, less resistance to disease, and ultimately shorter herdlife.    [In the case of their Guernsey herd, a lethal recessive hidden in the genes of the foundation sire multiplied to effectively breed the herd to extinction.]

At the same time, however, within certain closely bred lines, such as the Crescent Beauty Admirals (a Holstein strain developed by the Hetts families in Wisconsin), remarkable longevity with sustained health and fertility was achieved (perhaps at the expense of a slower rate of production maturity, which is why this bloodline lost popularity when "indexing" came into vogue).   Likewise, the Jersey breed, which has the highest average inbreeding coefficient of all widely dispersed dairy breeds, also has the longest average "Productive Life" as calculated from DHIA records.

So the question arises-- is inbreeding automatically bad, or is it a case of "it depends on which ancestors you choose"?         

For the dairymen of the 1950s, one of the attractions of AI was having a wider choice of bloodlines in available sires.     If your local bull-raising breeders were all in the "Burke" bloodline, over time your cows got shorter legged and deeper uddered-- but if you used a "Rag Apple" bull you could gain back some stature and some udder strength.     If your local bulls were all "Ormsby" blood, over time your cows may become narrower and frailer with more leg problems-- but if you used a "Crescent Beauty" or a "Carnation-Homestead" bull, you could get the width or strength back.   If your cows lacked adequate levels of bf% or longevity, a "Dunloggin" bull would breed it back in.   In the era of bloodlines and linebreeding, each line had a strength that was attractive to commercial dairymen's needs. 

Thus, when widespread young sire sampling began in the 1960s, pedigrees that were a mix of two or three bloodlines were more popular (and often proved out to be higher in rank) than the totally linebred young sires.

So the "linebred" herds of the 1940s, -50s, -60s, gave way to the "sire line" herds of the 1960s and -70s, when faster production and type progress was achieved.     Inbreeding fell out of favor, bloodlines became diluted by the shift in emphasis to "ranking sire" preferences in pedigrees.  Availability of sire proving data for production and type pushed pedigree selection aside.

Indexing-- a different (more insidious?) form of "inbreeding"

Sire evaluations kept evolving in the AI era, as breeding emphasis shifted.   Daughter-Dam comparisons gave way to daughter-herdmate comparisons, which then gave way to daughter- "modified contemporary" comparisons-- the first format in which contemporary "value" was adjusted by relative rankings of their sires.    By the 1980s, the system appeared to be running out of steam, with very few new sires exceeding the performance levels of established sires. 

At this point, two things happened concurrently.   The "Animal Model" method of comparing animals on all relationships-- ancestors, siblings, and progeny-- replaced the MCC as the basis of both sire and cow indexes.     The ranking of sires shifted from a multi trait selection (ie, what each dairyman individually considered important to him; type, or milk, or components, or some combination of all three) to "composite indexes"-- "TPI" for Holstein breeders, "PTI" for Jersey breeders, "Net Merit" for commercial dairymen.     In Holsteins, we developed composites for Udders (UDC) and Feet and legs (FLC) to "save time" over considering each individual trait.   The overriding emphasis became, ":How does this bull RANK?:"instead of, "What can this bull do for me in my herd?"

The Animal Model proved to be hidebound by Parent Averages [sire index + cow index / 2] and increasingly, we saw the same sires and maternal grandsires showing up in young sire pedigrees in all major AI systems-- an influence that eventually spread globally, given the choice of the dairymen in Western Europe to transform their dual purpose breeds into dairy "Holsteins" (who then competed for the semen on the highest protein yield sires in USA and Canada).   After a fairly short period of time, it was observed that pedigree inbreeding coefficients were rising in all major dairy breeds-- a consequence of the "same" sires being used in all YSP selections.

What was different this time was a recognition that negative trait expression was now occurring at levels of "inbreeding" much lower than was seen in the historical inbreeding trials.    DHI data began to show rising stillbirth rates, lower cow fertility, higher incidences of mastitis, and shorter herdlife-- in spite of a general market avoidance of "related" [sire x maternal grandsire] matings in daily practice.     In consequence, breed associations and AI studs both began to offer controls to reduce "inbreeding" in computer mating programs-- assuming the "problems" were the result of having common ancestors hidden behind the seemingly unrelated bulls and cows mated.

Is it "common ancestry", or is it "too narrow a gene pool"?

Common sense should have told us that, if we were seeing "inbreeding effects" in spite of the general attitude of avoiding related sire matings, is the problem really ":inbreeding:"-- or is it in fact, being too focused on too narrow a pattern of gene expression?

Concurrent with the adoption of "indexing" to "rank" individual sire and cow performance, was an increased preference for the taller, more angular, often narrow/shallow cow physique.    This was a selection based upon 1960s/early -70s type studies, which identified that physique with the higher producing first lactation heifer.    Keep in mind, in that era, University research herds were tie stall housed, individually "challenge" fed, and were under no requirement for calving interval.   In that same era, most feed nutrition research was based on seeing how much corn a cow would eat without gaining weight or getting sick.    The angular cow made milk instead of weight gain, and the higher milking of them generally avoided getting sick.   So major changes were made in our type preferences-- away from the "hay burner" physique, in favor of the "grain burner".

These changes found little favor with the older generation of breeders, whose preference for the wider, deeper, stronger cow physique was based in a recognition that "longevity" generally came in the wider, deeper, stronger-chested, fuller-boned package.    But high heifer production won out over longevity, and the type preferences were changed without regard for any expectation that at any future time we might wish to use less grain and concentrates than was pushing production up rapidly in the 1970s and -80s.

The external physique represents the genotypic balance

In the earlier AI eras, when sires were not "ranked" by composite indexes and bloodline selection was still a market sector, sire analysts maintained wide pedigree variety.    In Holsteins as noted you had Ormsby, Burke, Dunloggin, Rag Apple, Crescent Beauty, and Homestead major lines.  In tandem you had a variety of phenotypic expression-- tall vs wide, angular vs conditioned, strong and large vs dairy and refined.     You could mix them at will, and dairymen did-- and their herds maintained physical vigor as they gained in productivity.

    The "aAa" Breeding Guide (also known as Weeks' Analysis) grew to prominence in this era, by identifying and successfully blending these differing qualities through matings, helping cattle to be more productive over longer lifetimes-- maintaining heterosis effects in the matings, while limiting extreme physical expression in the results.           

When sire pedigree lines began to narrow, and physiques became more similarly oriented toward the higher levels of angularity, focused on ever-higher production volumes, systems like "aAa" adapted more precision, in an attempt to avoid the consequences of too high a level of gene homozygosity in available mating choices.    The best cattle became even more productive     

But within herds that bred randomly, totally on ranking index, or by computer pedigree screening, you would think we were back in the 1950s, with a constant 40% of heifers unable to return to second or later lactation production-- in this case lacking the strength or balance of physique to have a full functional lifetime.     High production Holstein dairymen in frustration embraced crossbreeding, using Jersey sires on heifers to insure calving survival, using "dual purpose" sires (Brown Swiss, Swedish Red, Normande, Montbeliarde, Fleckvieh) on cows to gain back some vigor.

And, in the near term, it worked-- justifying a belief that somehow, purebreeding was in fact unavoidably tainted by "inbreeding".     This represents a fundamental misunderstanbding of what "inbreeding" really is-- and what was really happening to those herds.

"Selection Depression"-- the result of single trait selection

In reality, what has destroyed profitability of commercial purebreds is the pop psychology of AI genetic interpretation.     Research had long since proved that "single trait" selection, while maximizing short term progress in the focus trait, produces longer term losses in most if not all "vigor" traits-- for the simple reason that the most extreme performer is usually not healthy or capble of sustained performance.     For the same reason that many athletes die at younger ages, higher production cows put too much of their substance into maximizing production, too little into being able to survive the nutrient strain of that production.

Tying elite heifer performance to the more angular physique was a recipe for disaster.   Prior to that industry decision, it did not matter to sire analysts what a high production cow looked like-- if she had proven superior performance, her type was acceptable.     Superiority was a decision based upon in-herd actual (realized) performance, not a pedigree-based, predicted (imputed) performance.

Dutch researchers in the 1990s, studying the effect of a mere three generations of North American sire usage on their native dual-purpose Friesians, found "selection depression" in those same vigor traits (normally associated with "inbreeding") was showing up in their herds-- harder first calvings, more stillbirths, slower rebreeding, more mastitis, shorter herdlife-- this in spite of not reaching the 8.25% inbreeding coefficient levels USA studies insisted was the usual threshold for observing "inbreeding depression".     The coined the term ":selection depression:" as a more accurate interpretation of what went wrong-- they had used the absolute highest protein yield sires, without regard to type, SCC scores, fertility rates, calving ease, ie, any exploration or estimation of health-related traits (for which we now summarize sire data).    In effect, another proof that single trait selection focus loses as much in what we do not consider important as it gains us in the one thing we chose to be important.   

The innate pairing of genes during conception is a symbolic representaiton of what nature has defined as a species adaptation and survival mechanism.    In selection, if you wish animals to perform for a healthy lifetime, in all desired functions, we need to "pair" performance traits with stamina traits; we need to balance the way the cow's system rations nutrient energy to all energy driven functions-- milk volume, solids percentage, reproduction, health maintenance (related to body conditioning).     We need to match "form" to "function".   We need to appreciate the more adaptable physiques over those possessing a handful of priority traits to a higher degree.

Will Genomic testing help or hinder avoidance of inbreeding?

Genomics has great potential to answer the "inbreeding" question in a biological, as opposed to a pedigree, sense.     This is due to Genomic ability to look directly at the DNA to see which genes are actually present (rather than to assume they are there from recorded ancestry).

Under pedigree-driven Parent Averaging, all full siblings have the "same"inbreeding coefficients.  Yet when you look at evaluation results for full brothers, we often see great differences in their trait expression.     Thus logic again tells us, these full brothers inherited a different mix of genes from the same parents.     In fact, high school biology already taught us that each conception is a unique, unduplicated individual.     It takes identical twinning or cloning to come close to a full sharing of genetic material-- even then minor sibling differences are expressed, as the genes possessed respond to the variables of the environment (which began in the womb).

Under Genomic testing, we now see the actual levle of homozygous gene pairings possessed in the DNA of individuals.    Thus we can calculate different levels of "inbreeding" for full siblings.   But note the major implication in this:  inbreeding is not pedigree linked.    The assumptions on which all our prior avoidance of mating related sires, or using computer mating pedigree models, are wrong.    "Ancestors in common" does not cause inbreeding depression-- any more than it could explain selection depression.     Too many homozygous gene pairings covering traits that must be antagonistic to fertility, calf vigor, disease resistance, even growth rate, is the true cause of "selection depression" (or what we have mislabelled "inbreeding depression").

This is not to say that too close "inbreeding" is not capable of bringing out of the closet some if not all the negative traits possessed in the gene makeup of mated animals.    Relationship makes possession of genes in common more likely.   But this applies to good as well as bad genes.     The best method of avoidance is a structural approach to mating when levels of relationship exist between mates-- remembering that the physical proportions and qualities possessed by cows and bulls express gene-driven patterns, thus can be quides to a desirable heterosis response from the mating effect.

 Genomic testing actually has the ability to make "selection depression" occur more frequently.  As you increase the number of matings in which both mates (sire and dam) possess the same "marker" genes associated with desired trait expressions (the narrow range of traits we have currently decided are linearly measurable and worth measuring), the level of homozygous pairings is bound to increase.     The "genetic effect" of selection focus on a few traits, can work at cross purposes to the "mating effect" of seeking the well-balanced, more environmentally adaptable physique.     Thus we will still see animals whose performance cannot live up to expectations that were based purely in possession of certain genes.    But we may also see the more expensive problem of animals who try to live up to expectations, accomplish it briefly, but fail due to the loss of "vigor"-- related to possession of substance and stamina qualities.