What is the real meaning and relevance of “cow family” selection?

This is from the November 2011 Dairy Route Letter

The more experienced purebred Breeders, those who thought about bloodlines prior to “index rank”, continue to seek out cow family sires--  bulls who come from developed maternal lines, often showing more longevity (thus higher fertility) than the conventional “sire stack” AI offerings. 

These Breeders recognize that “performance traits” and “maternal traits” rarely coexist in any individual sire, as underlying genes tend to be negatively correlated.     To produce cows who reproduce regularly and stay healthy over full lifetimes, but stay competitive for milk production, requires a “blending” of pedigrees, crossing “performance” sires on “maternal trait” cows, and vice versa, to keep some balance in their breed gene pool and maintain potential for heterosis in matings (“hybrid vigor”).

Single trait [rank]  vs  Multi trait [matrix]  selection  approaches

In other agricultural species [both plant and animal] parent stock lines are maintained with specific gene focuses, as in (a) growth rate, (b) fertility rate and litter survival, (c) milking ability, (d) fleshing ability.

In these species, there is no “one size fits all” composite genetic rank—there is instead a preservation of linebred stock lines, crossed to produce higher performance hybrid offspring.    In this way, plant as well as animal breeders maintain gene pool diversity, while improving yields.

The single trait selection approach (which began with PD Fat in the 1960s, PD Milk in the 1970s, then a move to PTA Protein in the 1980s, TPI in the 1990s, $Net Merit in the 2000s, Genomics in 2010) is thus peculiar to the mainstream dairy selection practices.     The fallacy in single-trait approaches (whether a single measured “trait” or a composite index “rank” is used) is that it leads us to make too many “like to like” matings—matings that intensify gene possession for a narrowly defined performance dependent on an equally narrow “ideal” environment.    These matings often lead to “inbreeding depression”.

The fallacy of composite index ranking over time

Under any single-trait approach, the more you accelerate yield gains in theory, the more you lose in the realized “secondary” (or ignored) supporting traits.   Traits given the most weight in the index will prove to dominate the resulting genotypes such that, in the third generation, most of the “heterosis” effect from initial switch to the selection index is gone.  Mediocre performance in all “secondary” traits is what now shows up in our cows, whether it is stalled production response (when the index favors health traits) or lost fertility, increased stillbirths, and lower immune function (when the index favors yield traits).

The concept of “performance” lines and “maternal trait” lines

In a practical sense, for dairy breeding today (given the huge loss in bloodline diversity that followed the pedigree-based approach to sire and cow ranking under “Animal Model” evaluation formulas) the better approach to both gain production and also gain health and reproduction is to alternate generations – if you have been following “performance” [sire stack] sire lines, switch for a generation or two in favor of  “cow family” [maternal line depth] sire lines.

In Holstein breeding history, this was done when the “Burke” [performance] line cows got too small and short lived (from udder failure) was crossed with the “Rag Apple” [maternal] line bulls, and the result—

gains in stature and strength, more mature production volume, longer breeding life—were exceeding the predictions from the individual sire line indexes.     Of course, the most dramatic example of this blood cross was  Round Oak Rag Apple Elevation  (inbred “Burke” sire from linebred “Rag Apple” dam).

Examples  of  “performance” sires  vs  “maternal line” sires

Two of the more extreme “performance” sires bred just prior to the introduction of “fitness” traits were Marathon BW Marshall and Stouder Morty.     Both of these sires were introduced to Holstein breeder circles as elite “production yield” sires with desirable “linear type traits”…     Both are now recognized as among the breed’s recent worst for calf stillbirths, daughter fertility, and thus productive life.   

They had some things in common.    First, they had dams that made a single huge lactation, but over short lifetimes had few calves and inconsistent breeding records.    Often such cows leave no maternal descendants to start a “cow family” due to the lack of (a) fertility and (b) maternal instinct.    A closer look at their pedigrees shows that each was from a line of cows with indifferent or inconsistent records.

These cows came to AI attention because they had high indexes, a result of having a high performance “sire stack” that was validated by their above-average first lactation yields.    But a sire-based “index” system has this weakness of understating differences in maternal traits of importance:  beginning with a trainable intelligence, followed by good natural fertility, then some maternal instinct at calving, an easy transition to group milking, a strong appetite, and good health consistent with a will to live.    Cows who come from a maternal trait inheritance will produce more live calves in their lifetime, and you will get more total salable milk production in their lifetime, than from a pure hybrid “performance” sire stack.

The  possible  weakness  in  our  initial  approach  to  Genomics

The accumulating defect in Genomic selection (as focused on TPI, $NM or LPI) is that all the emphasis is on the limited range of measured traits, that lean heavily toward “performance” over “maternal” cow phenotypes.    This is a consequence of basing all Genomic interpretation of females on a “sire” base—an error already implicit in “Animal Model” that contributed to losses in fertility, health and longevity.

In the USA, 70% of all young ranked Genomic sires have aAa’s of 2-3-4 combinations—clearly more “performance” than “maternal”.    In Canada under the LPI index, 70% of all ranked Genomic sires are 1-3-2 aAa combinations (similar to the US, but with less “strength” due to LPI’s lower health emphasis).

This is one way of observing that, rather than finding us “outlier” sires as predicted, Genomics to date is accelerating the process of “likes to likes” which leads to inbreeding depression.

Linear type and production yield measures are both “performance” in focus

Do not make the mistake of believing that “if I select high type I will avoid short herdlife cows”.   When you critically analyze that statement, common sense tells you that higher youthful type scores are just a prediction of earlier production maturity.     The way to add longevity to your herd is to look at the cow line in a bull’s pedigree and find cows like those behind 76H 466  Ridgedale Escalate, where you have three dams in a row that scored “EX 95” thus are proven to develop into maturity, and on their lifetime production performance, express an elite level of sustained productivity that requires successful fertility and health quality to realize.     (Pedigree may be an old tool, but it still contains relevant information.)

The stated genetic evaluations of sires like “Escalate” are never as high as you find from purely hybrid “performance” sire stacks.     But some clues to “maternal quality” are in the better ratings for stillbirths, daughter calving ease, daughter pregnancy rate, somatic cell scores, and Productive Life, verified from a maternal pedigree development that explains the gene source of that superiority in those traits.

These sire differences have always existed.     But who ever explained it to you this way?

Cow family sires      What goes around, comes back again

In the seminal days of AI, while many studs talked about “proven sires”, other studs (like Curtiss, our original home in the AI business) talked about “cow families”.     Comparing herds who followed each philosophy, you could find elite production in either group, and the same rate of production gains over time.     But this parallel progress was ignored by researchers who focused, not on total herd performance, but individual sire deviations.

When all we measured was lactation production and type scores, emphasizing one over the other, a great deal of bloodline diversity was lost.    Since the introduction of traits for health, fertility and “will to live”, it is much clearer that we were willing to trade off cow longevity in favor of rapid production realization and equally rapid cow obsolescence.

The “cow family” concept makes sense once you rename it “maternal traits” and consider the value in daily husbandry of the self-reliant cow who has a strong mothering instinct and equally strong will to live, attributes biologists see as linked to fertility.

Mich Livestock Service Inc--“For the Best in Bulls”, including knowledge to succeed