Imagine you want to buy a horse. But you’re a tall guy, so you want a tall horse. In your search, you find advertisements for two likely potentials. They both look like great ranch horses, but one is listed as 72 tall while the other is described as 132 tall.
If you chose based on those numbers alone, without knowing the measurement systems being used in each case, you would be short-changed if you chose the second horse because of its seemingly larger height number.
If you knew the measurement system each number was coming from—inches and centimeters, respectively—and knew how to convert those systems to make them comparable, it would be clear that the first horse is taller despite the seemingly lower number (72 in. versus 132 cm, a.k.a. 52 in.). If both horses were measured using the same system—let’s say, hands, making them 18 hands versus 13 hands—to begin with, it would make finding the taller horse even easier.
Expected progeny differences (EPDs) in the cattle world are a lot like our hypothetical tall horses problem. Most breeds’ EPDs are on different bases. An EPD basis is analogous to a measurement system; it’s the system in which their measurement numbers (i.e., EPDs) make sense. If you try to directly compare the weaning weight EPD on an Angus bull to the weaning weight EPD on a Hereford bull, for example, your success at getting a bull that meets your genetic goals will be about as successful as if you had chosen the second “tall” horse.
What is an EPD Basis?
As mentioned, the basis for a set of EPDs is like a measurement system; the basis is the system in which EPD numbers make sense. How that basis is set is fairly arbitrary, however.
“The choice of a base has absolutely nothing to do with science,” explained Dr. Matt Spangler, beef genetics specialist and professor at the University of Nebraska-Lincoln. To set an EPD basis, “A breed association might choose a point in time like a given year, or maybe it’s a reference of high-accuracy sires.”
Dr. Larry Kuehn—research lead in the Genetics, Breeding, and Animal Health Research Unit at the U.S. Meat Animal Research Center (US MARC) in Clay Center, NE—echoed this, saying, “The key take-home message is [different breeds’ EPDs] are on different bases because breed associations are working with their own data.”
He added that, with a few recent exceptions, breed associations don’t come up with their EPD bases with the ease of across-breed comparisons in mind—the breeds working with International Genetic Solutions being the notable exception. Mostly, a breed’s EPD basis is a quirk of their recordkeeping, data, and history.
There are some potential advertising benefits to a breed’s EPD basis. For instance, let’s say the natural distribution for an EPD has half the cattle with a negative EPD. If it is generally considered that higher is better for the trait, a breed can simply add a set number to every animals’ EPD, so that virtually all the animals have positive genetic predictions for the trait.
“If your breed’s EPDs for growth are so much larger than mine simply because your base is bigger, you can advertise better than I can,” Spangler explained. “That’s a large part about the base; how does it make me look compared to my competition?”
That advertising advantage only works if producers looking at the different breeds’ EPDs are unaware of the importance of the difference of EPD bases. This comes down to producer education.
“That is why Extension personnel, for a long time, have encouraged producers that—if they are comparing bulls of different breeds—they need to realize they are inherently not compatible, and they have to use some tool to make them comparable. Otherwise you might think that two bulls have exactly the same EPD, but—once you put them on the same base—they are vastly different.”
Converting EPD Bases with the Across-Breed Adjustment Factors
There are two main ways to compare numbers coming from different measurement systems; either convert them or start with the same measurement system in the first place. With the horse example, you can either convert centimeters into inches (2.54 cm = 1 in.) to get comparable heights, or you can measure the horses in the common system of hands to begin with.
When it comes to EPDs, the effort to “convert” EPDs using different bases came first in the form of the US MARC’s across-breed EPD adjustment factors. The idea for this tool came out of Beef Improvement Federation (BIF) annual meetings and the growing recognition of the value of crossbreeding and heterosis to the beef industry.
At the 1992 BIF meeting, Dr. Jim Brink of Colorado State University made the motion to have breed comparisons presented at the 1993 meeting. US MARC’s Dr. Larry Cundiff worked with Dr. Dave Notter of Virginia Tech, using data from US MARC’s decades-running germ plasm evaluation studies to come up with the statistical procedures to calculate the EPD base adjustments for the breeds in the US MARC database. Notter presented the first release of the adjustments at the 1993 BIF meeting. Updated across-breed EPD adjustment factors have been released annually since then in the annual BIF proceedings.
After that first year, US MARC scientists Drs. Cundiff and Dale Van Vleck took up the task of annually producing the adjustment tables. At first, adjustment were only available for growth traits, but carcass traits were added in 2009.
In the 2000s, Drs. Kuehn and Mark Thallman joined the US MARC staff, and with the impending retirement of Cundiff and Van Vleck, the calculation of the across-breed adjustments were transitioned to their oversight.
Spangler characterized the across-breed EPD adjustment factors as tools ranchers can use in making genetic decisions in a crossbreeding system.
“They’re really a set of additive adjustment factors that can be used to take EPDs from disparate genetic evaluations or breed associations and puts them on the same base and allows you to compare them,” he said.
“If you have an EPD from Angus and Hereford, for example, they aren’t directly comparable. The factors that US MARC puts out annually can be used to make those comparable.”
The adjustment factors were quite controversial when they came out. Breeds worried they wouldn’t look good. The American Angus Association was reported to be very concerned when Angus was made the base breed, worrying that it would not put their breed in the best light.
Over time, these concerns have gone away. The adjustments have improved greatly over the years, especially since US MARC started their continuous evaluation of breeds in 2003. The adjustments themselves put the EPDs on the same base, which is Angus.
This improvement doesn’t mean the adjustment factors are free of potential problems. Between various breeds’ genetic evaluations there are potential scaling differences, which means the units the EPDs are presented in by the individual breeds may vary slightly in magnitude. This results in some potential bias when evaluating EPDs from different genetic analyses, the farther they deviate from the mean.
Linearity—different numbers within a specific EPD being proportional in their magnitude—is also assumed between breeds’ EPDs, but that is not necessarily the case between the various analyses. Therefore, as the breeds’ EPDs move to the high and low extremes, they might not be exactly comparable. This means linearity is another area of potential bias on the tails of the EPDs.
It has always been known that the US MARC adjustments do a good job of adjusting the bases between different EPD systems, and these two potential problems with bias related to scaling and linearity have become less of a concern over the years with US MARC doing the long-term continuous evaluation of the most impactful 17 breeds.
Creating Truly Multibreed EPDs to Directly Compare Different Breeds
The effort to get cattle proverbially measured with the same measuring stick—with true multibreed EPDs that share a common basis—started later than the effort to convert existing EPDs. This lag was because creating true multibreed EPDs requires evaluations that don’t happen naturally in the industry very often.
“In order to get animals on the same base, they need to be compared to one another,” explained Kuehn.
“In general, animals from one breed are generally not compared in a contemporary group to animals from another breed. To get them on the same basis, they need to be compared to each other in common, contemporary groups.”
The first effort to develop a true multibreed EPD model came out of Cornell University and the American Simmental Association (ASA). Cornell scientists Drs. John Pollak, Dick Quaas, and postdoc Bert Klie developed the multibreed model in 1996, with the first production run in 1997.
There were a lot of technical obstacles to produce a multibreed genetic analysis. Founder effects of the base breeds used as the foundation for breeding up the newly imported Continental breeds had to be accounted for. Also, the heterogeneous variance from cattle of different percentages, as well as the varying heritabilities of breeds, needed to be taken into account. Incorporating external EPDs from popular breeds for crossbreeding—Angus being the primary example—whose associations do not directly participate in the analysis also presented a problem. Quaas tackled this and came up with methodology to use externals in 1999.
In 2010, ASA and the Red Angus Association of America formed a genetic analysis company that would become International Genetic Solutions (IGS). This company’s express purpose was to produce multibreed genetic predictions on the same base and scale.
When the new company was announced at the 2010 NCBA Convention, it was met with rave reviews. Commercial producers had long desired genetic predictions that would be directly comparable when designing crossbreeding systems.
The key item in this effort was developing a single base for all the breeds’ EPDs. In addition to being difficult to do, this was a political football because it required cooperation between breed associations and every breed wanted to be seen in the best possible light.
“A shared base requires no arithmetic—it’s just a group of breeders and breed association personnel sitting down and saying, ‘Alright, what do we want this base to be?’” explained Spangler.
“If you’re a Limousin producer, you’ve always been accustomed to EPDs for weaning weight kind of looking like this, in this ballpark. But if you’re from Red Angus, they always looked a little different because they were in a different ballpark. So, the discussion again for a base is nothing to do with science. It’s more we all need to meet in the middle and get adjusted to slightly different-looking values. It is what are we going to agree with in terms of a base. That’s all there is to it.”
IGS’s technical advisory committee took on the task, however, and came out with the common base in 2012 in large part due to the leadership of seedstock producer Donnell Brown from Throckmorton, TX.
With the system functioning, many breeds quickly signed up for IGS’ analysis from 12 breed associations including Simmental, Red Angus, Gelbvieh, Limousin, Shorthorn, Chianina, and Maine-Anjou, as well as most of their Canadian counterparts. This unprecedented cooperative effort provides powerful objective genetic selection tools to compare within breeds and between breeds when designing crossbreeding programs.
The quality of the across-breed comparisons between the participating breeds was determined by how closely tied the databases were. This means each database had to have common sires’ calves represented in contemporary groups along with sires’ calves from the given breed. This primarily is Angus tying the datasets together through the various breeds’ hybrid registries with Red Angus being the other important breed. However, these relationships through common sires tied some breeds together better than others.
In 2018 when IGS switched to a new single-step model powered by BOLT software, how tied the breeds were came under tighter scrutiny. At the beginning, the feeling was they might not be adequately tied, so utilizing US MARC adjustments in IGS’ analysis was considered to make the genetic predictions more comparable. However, when the datasets were scrubbed for use in the new automated system to run genetic predictions weekly, it was discovered that they were actually very well tied for most traits. Only the carcass trait comparisons were enhanced with the US MARC adjustments embedded into the IGS analysis.
Overall, IGS is producing the highest quality multibreed EPDs available in the industry, with a database approaching 18 million head, making it the largest across-breed dataset in the world. This allows commercial producers the ability to directly compare individual animals from different participating breeds in order to design the best possible crossbreeding system, one that makes optimum use of breed complementarity and heterosis.
The IGS analysis will continue to improve in the years to come. In addition to the current list of directly comparable genetic predictions, IGS is working furiously to expand the number of multi-breed traits analyzed. These include heifer pregnancy, intake, days on feed, feet and leg structure, disease susceptibility, and mature weight. When completed, this expanded list of traits will allow managers to fine-tune their crossbreeding programs as never before.
Using These Powerful Tools
If you currently use or hope to include crossbreeding in your herd, US MARC’s across-breed EPD adjustment factors and the IGS multibreed EPDs are powerful tools in your genetic toolbox.
For instance, if a bull from breed A and one from breed B are both from associations participating in the IGS genetic analysis, the relative difference is simply A minus B. Managers comparing animals this way can be assured the EPDs are on the same base and scale, and the differences are linear.
If one or both breeds are not in the IGS analysis, the US MARC adjustment factors must be applied to put them on an Angus base. For instance, if the adjustment factor for A’s breed for yearling weight is -35 and the adjustment factor for B’s breed is -40, their direct comparison on an Angus base is [(A – 35) – (B – 40)].
It is important to note that the comparisons between animals from breeds directly from the IGS analysis will likely be different than when the same animals are compared using the US MARC adjustment factors putting them on an Angus base.
For most traits, the differences from the IGS analysis were obtained from the extremely large multibreed field database as well as their genomics. In contrast, the adjustment factors to put the animals on the same base from US MARC were developed from a relatively small but high-quality database obtained from their experimental herd. These tightly controlled experimental matings do an excellent job of characterizing the relatively few influential sires used from each breed within the US MARC herd, which are then extrapolated to breed-wide adjustments.
Both do a good job of comparing animals, but if the animals are in IGS’ pooled analysis, it is best to do direct comparison, saving the US MARC adjustments to compare breeds with Angus or others that are not in the IGS analysis.
Since most crossbreeding systems involve the Angus breed—which does an Angus-only genetic analysis—the US MARC across-breed adjustments will remain critically important to breeders designing an Angus-based crossbreeding system. Red Angus is the other common base breed in crossbreeding, and its inclusion in the IGS analysis makes it easy to build a crossbreeding system with the other participating breeds.
Spangler put this concept quite simply, saying, “For producers who utilize multiple breeds of bulls in a well-structured breeding program, they have to compare bulls from different breeds. So, having a common, shared base—or using tools like the across-breed EPD adjustment factors—becomes very important.”