This item is only available as the following downloads:
AN270 The Use and Economic Value of the 3K SNP Genomic Test for Calves on Dairy Farms1Albert De Vries, David T. Galligan, and John B. Cole2 1. This document is AN270, one of a series of the Department of Animal Sciences, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida. Original publication date September 2011. Visit the EDIS website at http://edis.ifas.u.edu 2. Albert De Vries, associate professor, Department of Animal Sciences, University of Florida; David T. Galligan, professor, School of Veterinary Medicine, University of Pennsylvania, Kennett Square, PA; John B. Cole, research geneticist, Animal Improvement Programs Laboratory, ARS, USDA, Beltsville, MD; Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL 32611.The Institute of Food and Agricultural Sciences (IFAS) is an Equal Opportunity Institution authorized to provide research, educational information and other services only to individuals and institutions that function with non-discrimination with respect to race, creed, color, religion, age, disability, sex, sexual orientation, marital status, national origin, political opinions or aliations. U.S. Department of Agriculture, Cooperative Extension Service, University of Florida, IFAS, Florida A&M University Cooperative Extension Program, and Boards of County Commissioners Cooperating. Millie Ferrer-Chancy, Interim DeanDairy producers have had the opportunity to test their female animals with the low density 3K SNP genomic test since September 2010. e 3K genomic test provides an estimate of an animals genetic merit for many traits, including milk production and Net Merit (NM$). As one of several available genomic tests, the 3K genomic test works by comparing an animals DNA to a database that associates DNA patterns with genetic merits of traits. erefore, a genomic test can provide a fairly accurate estimate of an animals genetic merit early in her life without any other data, such as her phenotypic records or information from parents or siblings. Various vendors (for example, Holstein Association USA and Pzer Animal Health) sell genomic test kits that help a producer collect a DNA sample and send it to a processing oce. e USDA then calculates the genetic merits of the traits of the animal tested. e producer gets the results back within a month or two. As of August 2011, approximately 45,000 animals have been tested with the 3K genomic test, most of them females. Still, many dairy producers wonder if the 3K genomic test might have value for their operation. e benets of using a 3K genomic test include discovering or conrming parentage for mating decisions that minimize inbreeding and selecting candidates for embryo transfer. Our objective in this article is to explore how dairy produc ers who primarily sell milk might benet from using a 3K genomic test on young calves in order to select which calves to raise as replacements. Non-selected surplus calves would then be sold at an early age. Because of the increase in reproductive eciency and use of sexed semen is producing heifer calves on many dairy farms, choosing which calves to raise based on their genetic merit, among other factors such as early life health events, has become a real option that needs to be considered. Later in the publication, well also briey address the topic of how many, if any, heifer calves can be considered surplus. Genetic progress is made by selecting superior animals as the parents of future generations. If all heifer calves are raised, virtually no genetic progress is made on the female side. In other words, all genetic progress in the herd then comes only from using genetically superior AI (articial insemination) sires. But if there is a way to select the genetically better heifer calves to be raised as replacements, the dairy producer can make genetic progress on the female side as well, which in turn causes total genetic progress to increase faster. One of the values the 3K genomic test provides is an animals genetic merit for NM$. Net Merit is an estimate of the expected lifetime prot of a female compared to the breed base (an average cow born in 2005) in the same environment; this trait shows a direct impact on the income an animal can generate within its lifetime and later aects its ospring. e NM$ index includes economically relevant traits related to milk yield, health, longevity, fertility, calving ease, etc. An animals breeding value is her genetic merit
2compared to the genetic merit of the breed base animal. For example, a calf with a breeding value of $300 for NM$ is expected to be $450 more protable during her productive life (about 3 lactations) than a calf with a NM$ of -$150, provided that all environmental factors are the same. Selecting the calves with the highest NM$ should directly impact the protability of lactating cows. Furthermore, the daughters and future generations of the selected $300 NM$ calf are expected to have a greater NM$ (in a decreasing way) than the future generations of the calf with the -$150 NM$ genetic merit.What Is Needed for Genetic Progress?ree factors determine the amount of genetic progress made in one generation in a population (for example, the current group of available heifer calves). First, there must be genetic variation in the trait NM$ in the population of calves. is variation is expressed by the standard deviation. Estimates of the standard deviation of the breeding value of NM$ vary from approximately $300 to $400. In the analysis below we chose a standard deviation of $350. If the standard deviation is $0, that would mean that genetically all animals are the same and no superior animals can be selected regardless of how good the genomic test is. Second, genetic progress depends on how accurately we can estimate the true breeding value of an animal for NM$. is true breeding value is unknown, but a 3K genomic test provides a good estimate of that true breeding value with a reliability of approximately 65%. If only the sire of the calf is identied (meaning there is no genomic test information), the reliability of her breeding value for NM$ would be about 20%. If a calfs full pedigree is identied, the reli ability of her breeding value for NM$ would be about 34%. ese traditional methods of estimating a calfs breeding value for NM$ have a lower percentage of reliability than that obtained from a genomic test because in these traditional methods it is not known which sample of the good or bad genes the calf inherited by chance from her parents. e reliability is also a measure of how well we can rank animals on their true breeding values based on a prediction of those breeding values (for example, if the information is provided by the 3K genomic test or the information is from relatives). us, a 3K genomic test results in a better ranking of calves on NM$ breeding values, among other traits. e third component of genetic progress is the selection intensity. is is a function of the fraction of ranked animals actually selected. e fewer calves selected, the greater the selection intensity. If the top 90% of the calves are selected (almost all), the average breeding value will be lower than if the top 50% of the calves are selected. In other words, the smaller the fraction selected, the greater the average breeding value of the selected animals. On most commercial dairy farms, the supply of heifer calves will not be much greater than the number needed to be raised as replacement animals. erefore, the selection intensity is low; perhaps only 10% to 30% of heifer calves could be called surplus and culled. Purchasing a 3K genomic test for a heifer calf is an investment. e cost of a 3K genomic test is approximately $40 per animal. Now the question is clear: Is there value in using a 3K genomic test in order to rank animals better by their NM$ breeding values and to increase genetic progress, given that only a certain number of heifer calves need to be selected? More specically, which calves should be tested, and how does that depend on pre-ranking of calves based on traditional sire-only or full-pedigree information? In order to be protable, the value of using the selected calves to increase the genetic progress must exceed the cost of testing the calves. In the analysis that follows, the average value of the kept calves depends on their estimated genetic values, whereas the average value of the calves sold does not depend on their estimated genetic values. We wrote a simulation program that tested various fractions of calves with the 3K genomic test (for example, all calves, the top 30% if calves were pre-ranked, the bottom 40%, the calves ranked 30% to 80%, etc.). e genetic progress of the kept calves, as well as the total cost of testing, and the net value of the test was calculated. For example, if 90% of all calves are tested, and 80% of all calves are kept, then the cost of testing per kept calf is $40 0.9 / 0.8 = $45. If the increase in genetic progress of the average kept calf is worth $100 as result of the testing, then the value of the test would be $100 $45 = $55 per kept calf.Value of the 3K Genomic Test When Calves Cannot Be Preranked on NM$ Breeding ValueWhen calves cannot be pre-ranked on NM$ breeding value, we assume that we have no information about a calfs genetic potential for milk production, fertility, longevity, etc. is may be the case when natural service bulls are the sires of the calves or no genetic information from the AI sires is available. Before testing, all calves are considered equal and the calves selected (kept) would be on average of
3the same genetic value as the calves not selected. Applying a 3K genomic test to some or all of these calves has the greatest value to a dairy producer, compared to a scenario in which pre-ranking is possible with a reliability > 0%. Table 1 shows the value of a $40 3K genomic test per kept calf, depending on how many calves are tested and how many of the available calves need to be kept. e table shows that all calves should be tested to obtain the greatest net value per selected calf. Testing more calves increased the average genetic value of the kept calves, as well as the cost of testing per kept calf. Yet, by testing more calves, the increase in genetic value is greater than the increase in the cost of testing. When all calves are tested, the value of the test per kept calf is $32, $87, or $137 depending on whether 90%, 80%, or 70% of the tested calves need to be kept. Using information from Table 1, it would be wrong to conclude that keeping fewer calves is more valuable than keeping more calves. Determining how many calves to keep as replacement heifers on a dairy farm requires a complicated analysis. For example, replacing more cows faster increases genetic progress and aects production of the current herd. Further, the availability of excellent reproduc tive programs and sexed semen allows dairy producers some exibility in how many surplus heifer calves they can create, so selection intensity could vary. Other sources of information that predict a calfs future performance should also be considered, including health events early in life, the dams age or calving diculty, or season of calving. Alterna tively, the expenses of the 3K genomic testing could be used instead of purchasing more expensive semen from sires with a greater genetic merit, or it could be used elsewhere if the money spent would result in a greater return on investment. If semen from more superior AI is purchased, genetic progress is then increased through the male side instead of the female side. Another option is embryo transfer from selected females, which makes determining the costs and benets of these dierent methods even more complicated. We are currently quantifying many of these aspects of this complicated but interesting problem. e goal is to provide dairy producers with some guidelines that take all important factors into consideration.Value of the 3K Genomic Test When Calves Can Be Pre-ranked with 20% ReliabilityCalves can be pre-ranked for genetic merit when the genetic merit of a relative (or relatives) is known, such as when their sire or full pedigree is known. Applying a 3K genomic test to such calves is less valuable because we can already rank these calves on genetic merit with some accuracy. Assume that all calves can be pre-ranked for breeding value of NM$ with a reliability of 20%, such as when their sire is identied. If the top 90% of calves are kept (without applying a 3K genomic test), the increase in breeding value of these selected calves compared to all calves is approximately $43. When the top 80% or 70% is kept, the advantage in breeding value for NM$ of the kept calves increases to $76 Table 1. Value of genetic superiority in Net Merit $ ( NM$), cost of 3K genomic testing (Cost), and net value of the 3K genomic test (Value), all per selected calf. Value = NM$ Cost. The dairy producer has earlier decided to select (keep) 90%, 80%, or 70% of all available calves. 90% selected 80% selected 70% selected 90% selected 80% selected 70% selected 90% selected 80% selected 70% selected Calves tested NM$ NM$ NM$ Cost, $ Cost, $ Cost, $ Value, $ Value, $ Value, $ 0% 0 0 0 0 0 0 0 0 0 10% 8 13 19 4 5 6 3 8 14 20% 15 28 39 9 10 11 7 18 28 30% 23 41 58 13 15 17 10 26 41 40% 30 54 77 18 20 23 13 35 55 50% 38 69 98 22 25 29 16 43 69 60% 46 82 117 27 30 34 19 53 83 70% 53 96 136 31 35 40 22 61 96 80% 61 110 156 36 40 46 25 70 110 90% 68 123 175 40 45 51 29 78 124 100% 76 136 194 44 50 57 32 87 137 Assumptions: Standard deviation of breeding values is $350 (multiplied by 1.39 to account for genetic progress in two future generations and 5% annual interest); cost per 3K genomic test is $40; and reliability of the 3K genomic test is 65%. There is no pre-ranking of calves.
4and $108, respectively. is gain comes from having only the traditional sire information available. We assumed no cost for the sire identication that gave the 20% reliability of the pre-ranking. Figure 1 shows the value of testing a fraction of these pre-ranked calves with a 3K genomic test. e gure also shows which range of calves to test. In this range, 0% is the highest pre-ranked calf for NM$, and 100% is the lowest pre-ranked calf for NM$. ese values are a combination of the increase in average breeding value and the increase in the cost of testing with the 3K genomic test when more calves are tested. Not all calves need to be tested with the 3K genomic test because calves that are pre-ranked high are very likely to be good enough to be selected. It does not pay to test them. Figure 1 primarily shows that calves ranked in the bottom 50% (pre-ranking 50% to 100%) should be tested. However, the range depends on the number of calves that needs to be kept. For example, if 90% of all calves need to be kept, the best policy is to test the bottom 30% (preranked 70% to 100%) of calves when they are pre-ranked with 20% reliability. e value of testing the bottom 30% with a 3K genomic test is $15 per kept calf. Testing other ranges (in increments of 10%) is less protable, although not by much. e fewer calves are kept (70% instead of 90%), the greater the value of testing. Furthermore, the optimal range of calves to test changes with the fraction of calves kept. Testing all calves increased the net value of the test per kept calf by -$10, $11, or $30 when 90%, 80%, or 70% of the calves were kept.Value of the 3K Genomic Test When Calves Can Be Pre-ranked with 34% ReliabilityNow assume that all calves can be pre-ranked for breeding value of NM$ with a reliability of 34%, such as when Figure 1. Value of the 3K genomic test per kept calf, depending on how many calves are kept, and the range of pre-ranked calves tested with the 3K genomic test. All calves are pre-ranked for breeding value of NM$ with 20% reliability. The 3K genomic test is applied to a fraction of the preranked calves (0% is the highest pre-ranked calf for NM$, and 100% is the lowest pre-ranked calf for NM$).
5genetic information on the sire and dam is available. Again, in the current analysis we assumed no cost to obtain the 34% reliability for the pre-ranking. If the top 90% of calves are kept (without applying a 3K genomic test), the average increase in breeding value of these selected calves compared to all calves is approximately $55. When the top 80% or 70% is kept, the advantage of the average breeding value for NM$ of the kept calves increases to $99 and $141, respectively. is gain from having traditional full pedigree information available is greater than when only the sire is identied. Testing calves with the 3K genomic test is less valuable when pre-ranking is done more accurately. Still, Figure 2 shows that testing the correct range of calves can make the 3K genomic test add value in addition to the pre-ranking. When 90% of calves are kept, at most $7 per kept calf can be gained. e bottom 30% of calves would be tested. Testing all calves decreased the net value of testing per kept calf by $24, $12, or $3 when 90%, 80%, or 70% of the calves were kept. erefore, testing all calves is not cost eective. In practice, the reliability of the predicted breeding values of the 3K genomic test results depends on the other information available. Calves with full pedigree informa tion would have a slightly higher reliability of the breeding values aer the 3K genomic test compared to when no prior information is available. However, this dierence is small. In this article, we used 65% reliability, regardless of the availability of other information. e accuracy of parent identication also plays a role. e availability of genomic tests is rapidly changing genetics in the dairy industry. AI companies have been using genomics to select AI sires. Now also commercial dairy producers can nd value in testing their calves to help decide which ones to keep. Figure 2. Value of the 3K genomic test per kept calf, depending on how many calves are kept and the range of pre-ranked calves tested with the 3K genomic test. All calves are pre-ranked for breeding value of NM$ with 34% reliability. The 3K genomic test is applied to a fraction of the preranked calves (0% is the highest pre-ranked calf for NM$, and 100% is the lowest pre-ranked calf for NM$).