During the past ten years, the use of Expected Progeny Differences (EPDs) by both seedstock producers and commercial cattlemen has increased markedly. EPDs have proven to be very valuable for discerning within breed sire differences, but are still of no value for comparing sires between breeds. Even though EPDs were never intended for across breed comparisons, many commercial cattlemen find this limitation to be very frustrating. Represented in Table 1 are the birth year EPD averages for all animals born in 1990 for six different breeds. One quick glance reveals why EPDs cannot be compared across breeds. Anyone with knowledge of breed differences will quickly recognize that Angus are not going to average 2.7 (3.1 - .4) pounds more than Simmental at birth. Furthermore, it would be incorrect to assume that Herefords would average nearly 30 pounds more at one year of age and produce significantly more milk (7.0 versus .6) than Simmental. These breed average figures help illustrate the confusion that can develop when bull buyers attempt to compare bulls of different breeds. Why do these EPDs vary so much? First and foremost, the breed populations are totally different and there are virtually no direct comparisons of sires among breeds represented in the breed databases. The only reliable breed comparison data resides primarily in university and government research projects. Secondly, the base reference point for each breed is different and thirdly, each breed has a unique genetic trend. Additionally, when crossbreeding systems were first designed, it was under the assumption that breeds would retain their distinct characteristics. However, this has not been the case as many breeds have changed resulting in increased similarities. This coupled with greater sophistication of many crossbreeding systems now being used has elevated the need for commercial producers to directly compare sires across breeds. What's more, it is common for breeders to maintain more than one breedto better accommodate their customer s bull needs. Consequently, the differences in EPDs between breeds can also be challenging for seedstock producers. Producers are increasingly developing "job descriptions" for their bulls and expect bulls within breeds to meet specific requirements. Any number of examples could be cited where producers may want to compare sires of different breeds in order to select bulls that more nearly match their specifications. Presently, the only way to compare sires across breeds is to develop a good level of understanding of breed differences based on well designed breed comparison research. Even with this knowledge, however, it would still be much less confusing if producers could directly compare EPDs. The question becomes, if there is such a strong demand for interbreed EPDs, then why don t we just go out and calculate them and be done with it. Unfortunately, it s not quite that simple. The purpose of this presentation will be to discuss a few of the challenges of generating accurate interbreed EPDs. NECESSARY INFORMATION Some of the necessary information needed to calculate accurate interbreed EPDs includes: Breed constants appropriate to the breeds of interest and to the environments and mating systems being considered. Heterosis adjustments which would potentially differ among crosses. Knowledge of the reference base (zero EPD point) for each breed. Sire EPDs appropriate for prediction of cross bred performance. Knowledge of possible genotype by environment interactions. BREED CONSTANTS Breed constants may be derived from breed comparison research or industry crossbreeding programs. Unfortunately, not all breed comparison research will produce good breed constants because of sampling of sires within the breeds being compared. Use of bulls in such experiments with accurate EPDs is preferred because the data can be adjusted for sire sampling. Unfortunately, many breed evaluation experiments have not used breed association EPDs as part of their analysis of breed differences. However, one of the most comprehensive breed comparison research projects that did use A.I. sires with EPDs is the Germ-Plasm Evaluation (GPE) Program at the U.S. Meat Animal Research Center (MARC). Early evaluations did not consider sire EPDs, however, a more recent analysis of the MARC data has included this information and should provide more accurate breed constants. Additionally, there are other research projects throughout the United States that could possibly be re-analyzed to generate breed constants. One in particular is Project NC-196, which is national in scope involving twenty cooperating research stations. Plans have been made to develop a database contributed to by all stations. This information will eventually make it possible to generate viable breed constants for use in the calculation of interbreed EPDs. HETEROSIS Heterosis is defined as the difference between the average of the reciprocal crosses and the average of the parental purebreeds contributing to the cross. Effects of heterosis in Bos taurus X Bos taurus breeds and in crosses of Bos indicus (Brahman type) X Bos taurus breeds are shown in Table 2. The values are average, for several experiments. The data indicate that heterosis effects vary with different traits and with different breed crosses. For example, heterosis is greater in Bos indicus X Bos taurus cattle in subtropical environments than Bos taurus X Bos taurus crosses in temperate regions in the U.S. This creates a challenge in the calculation of interbreed EPDs because of the different magnitudes of specific heterosis used to calculate the breed constants. Breed comparisons could be biased since they may not reflect the actual genetic differences among the breeds. Dr. L.D. Van Vleck concluded in his research that differences in specific heterosis as small as two to four percent could bias the breed constants enough to change breed rank. With this in mind, Dr. Van Vleck questioned the advisability of using crossbred data to generate breed EPD tables until more was known about specific heterosis. BREED BASES AND GENETIC TRENDS In order to connect within breed EPDs to breed differences, the reference points for each breed's EPDs must be defined. The reference point is sometimes technically defined as the population of foundation animals used to build up the relationships in the data set. EPDs are then expressed relative to the foundation animals that begin the accumulation of pedigree relationships. In most of the British breeds, this group dates back to the middle 1970's, whereas the base or zero point population for most Continental breeds is in the 1980's. Today the three kinds of bases being used by breed associations in North America are floating base, fixed base, and rolling base. A floating base is one in which the group of animals representing the base may change from one year to the next. If there is genetic trend in the population, then the zero point is different from one evaluation to the next. Consequently, a sire s EPD will change in proportion to the genetic trend. For example, if the trend for yearling weight averaged 1 lb. per year, a bull with an EPD of +20 in 1980 may actually be +10 by 1990. The bull is the same genetically, however, he has changed compared to the average of the base population. It is important to keep in mind that zero (0.0) is average for the base population, not the entire breed. If the base is fixed, genetic trend will not change the bull s EPD. For example, if a bull has a weaning weight EPD of +10 he will remain at + 10 with each new evaluation. An advantage of a fixed base is that a producer can set standards that will not change from one evaluation to the next. The advantage of a floating base is that EPDs keep up with genetic trend and there will not be a need to change the base reference point. The Canadian National Cattle Evaluation uses a three-year rolling base which is calculated by including all animals born in the current and previous two years. Instead of the base being fixed to a single birth year or population, the base is actually fixed to the most recent three year average. A summary of six breeds' zero reference points for birth weight, yearling weight and milk is given in Table 3. As you can see, the bases range from 1970 for Hereford milk to 1984 for Limousin birth weight and weaning weight. It has been suggested that all breeds should have a common base. Even though a common base is not necessary to calculate interbreed EPDs, the positive educational value was considered during a special symposium on interbreed EPDs, held in October1989. Dr. John Pollack, Cornell University, chaired a committee charged with developing a recommendation for a common base. The committee suggested during the May, 1990 Beef Improvement Federation (BIF) meeting that each breed fix its base at 1982. Breed associations with national cattle evaluation programs were asked to evaluate the impact that a base fixed at 1982 would have on their breed s EPDs. The results were presented at the May 1991 BIF meeting. Printed in Table 4 is an abbreviated summary of birth weight, yearling weight, and milk. As you can see, some breeds would be affected much more than others. Specifically, those breeds with bases in the 1970 s that have seen considerable genetic trend would be affected greater by a base move to 1982 than with those breeds with more current bases and less genetic trend. The May 1991 meeting was followed by a survey in which breed associations were asked to give their opinions about the fixed base proposal. Of the 12 breeds that responded, seven voted in favor of a base fixed at either 1982 or 1985. Not surprising, those breeds that voted against a fixed base were the ones that would likely be affected most by a base different than what was currently being used. It s interesting to note that nearly all of the breed representatives were concerned that a standardized fixed base would lead some people to erroneously conclude that EPDs could be compared across breeds. It is clear that getting all breeds to agree on a common base is not likely. But as stated earlier, this is not really a barrier since a common base is not a requirement for calculating interbreed EPDs. GENETIC X ENVIRONMENT INTERACTIONS After a thorough review of the research literature, Dr. Larry Cundiff concluded in a paper he presented at the 1989 BIF Convention that interbreed EPDs should be derived separately from experiments conducted in temperate and subtropical regions if Bos indicus breeds are to be compared to Bos taurus breeds. He went on to conclude that genotype X environment interactions are not important for weaning weight among Bos taurus breed crosses in temperate regions and that weaning weight EPDs across Bos taurus breeds could be derived from one or more experiments conducted under temperate conditions. However, analysis of data from more experiments may be required to accurately assess calving assistance. Even on a within breed basis, EPDs may have some shortcomings when comparing sires for use in temperate environments versus subtropical environments. Dr. Cundiff suggested that within breed EPDs for herds located in subtropical regions should be computed from herds located in subtropical regions and EPDs for herds in temperate regions should be computed for herds located in temperate regions. It does not appear, however, that sire X contemporary group interactions and sire X herd interactions are large enough to significantly reduce selection response. In summary, depending on the breeds, the traits and diversity of environments, some consideration to breed X environment interactions may be advisable when calculating interbreed EPDs. WOULDN'T THERE BE WINNERS AND LOSERS? This is a common question that, quite frankly, may be a misconception. If the focus was on a single trait such as growth, or milk production, or calving ease, this would be a valid concern. But, realistically, no breed is superior in all traits. Therefore, when interbreed EPDs become a reality, some breeds will excel in certain traits and be below average in other traits. There will be no winners and losers, but, hopefully, just a more accurate means of comparing sires regardless of the breed. This is why including more traits in the interbreed EPD analysis is important. For example, some breeds may excel in growth and milk, but without a good mature size EPD, it will be difficult to assess sires across breeds for efficiency. Other traits like calving ease, scrotal circumference, and lean yield should also be included in order to obtain a complete genetic picture. Astute commercial cattlemen will take advantage of the differences and use sires that best fit their needs. BREED ASSOCIATIONS' VIEWPOINT Given the significant investment in their respective databases, it is no wonder why some breeds are reluctant to unconditionally support interbreed EPDs. It has been estimated that breed associations and their members have collectively invested over $80 million in their breed databases. Considerably more money and effort have been committed to the development, calculation, and use of within breed EPDs. Moreover, breed associations have been dedicated to building the credibility and accuracy of EPDs. There is great concern that if interbreed EPDs fail to be as accurate as within breed EPDs, the credibility of the entire EPD concept could be jeopardized. Breed associations are not opposed to interbreed EPDs, but, their cautious approach is understandable. In fact, in January 1990, the U.S. Beef Breeds Council expressed their support of interbreed EPD research subject to: Greater efforts made to add to the database which would be used to develop these predictions. Research personnel, in cooperation with the Beef Improvement Federation, continue to examine the concept, evaluate the implications, and apprise the industry about the appropriate manner in which to interpret and utilize this information. That no data be released until all breeds who are members of the U.S. Beef Breeds Council with sire summaries be examined for utilization, application, and inclusion in the published reports. That BIF develop appropriate industry guidelines for the uniform application of a methodology to produce a cross-breed EPDs under the conditions cited in 1, 2, and 3. SUMMARY Given the potential value and interest in interbreed EPDs, they no doubt will become a reality within the next three to five years. More research will be required to generate the databases needed to produce representative breed constants. Other areas of question such as the influence of specific heterosis and breed X region interactions will also require attention. Provided adequate funding is available, the information and technology will be in place to produce viable interbreed EPDs. Their ultimate value will be determined by the end-users. In the meantime, it remains essential that producers: 1) conduct an audit of their resources; 2) determine what type of cattle will best match their resources and marketing goals; 3) develop a job description for the bulls of each breed to be used in the crossbreeding system; 4) have a good understanding of breed differences; and, 5) use within breed EPDs to identify bulls that best match their specifications.