| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ANIMAL GENETICS |
* Department of Animal Biotechnology, University of Nevada, Reno 89557;
and
University of Nevada Cooperative Extension Livestock Specialist, 701 Walnut, Elko, Nevada 89801; and
College of Agriculture, Biotechnology and Natural Resources, University of Nevada, Reno, NV 89557
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
Key Words: beef cattle • benefit-cost analysis • DNA marker • paternity test
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
The DNA technologies allow for paternity testing. The DNA markers of choice in paternity testing are usually microsatellites (Heyen et al., 1997
) that are codominant DNA markers (Fries et al., 1990), although SNP have also been proposed (Heaton et al., 2002). Essentially, the typing of several microsatellites is carried out in the offspring and in the alleged parent. A sire is eliminated as a parent when the genotype of the offspring is not compatible with the parental genotype for at least 1 microsatellite. As the probability of exclusion is the probability of rejecting an alleged parent that is a random individual within the population, the probability of exclusion depends on the marker type, the number of alleles, and the allele frequencies in the population to be used for paternity testing.
Economic analysis, which incorporates not only the benefits of increased calf performance, but also the cost of genotyping, should be investigated before implementation of a DNA paternity testing program. To our knowledge, there are no publications addressing the economic value of DNA paternity testing in free range beef cattle operations.
The objective of this paper was to propose an economic assessment of DNA paternity testing in beef cattle operations by a benefit-cost analysis. This analysis required computation of probabilities of exclusion for a set of DNA markers. We used 15 highly polymorphic microsatellites tested across 8 beef cattle ranches in the high desert of Nevada as an example of how this information can be incorporated into the cost-benefit analysis. Benefit-cost analysis with an incomplete DNA paternity identification program using a low number of microsatellites was also investigated.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
DNA Sources of Eight Nevada Beef Cattle Ranches
Ear notches were sampled from a total of 2,196 animals from 8 Nevada beef cattle ranches between 2001 and 2003. The number of sires, cows, heifers, and calves is depicted in Table 1. Samples were taken in the chute when cattle were palpated or vaccinated yearly. The Nevada beef cattle ranches chosen for this study were a representative sample of the northern Nevada beef cattle population. The ranches were named according to the northern Nevada county where they are located: 1) Humboldt1, 2) Humboldt2, 3) Humboldt3, 4) White Pine, 5) Eureka, 6) Churchill, 7) Washoe, and 8) Elko. All of these ranches have crossbred cattle with varied breed composition, with the exception of the Washoe Ranch, which is a purebred Angus operation.
|
Panels of microsatellites have been widely used for genetic diversity studies in beef cattle (Martin-Burriel et al., 1999; Kantanen et al., 2000; Edwards et al., 2000; Canon et al., 2001; Maudet et al., 2002; Rendo et al., 2004). For this study, 15 microsatellites were chosen from public sources based on their high polymorphism (number of alleles): BMC4228, BMS1226, BMS1244, BMS1315, BMS1634, BMS1789, BMS2055, BMS2573, BMS410, BMS499, BMS601, BMS650, ILSTS058, ILSTS081, and TGLA227 (http://www.marc.usda.gov/genome/cattle/cattle.html).
A total of 31,571 genotypes were determined for the 15 DNA microsatellites across the 8 ranches. The number of animals genotyped for each microsatellite varied between 1,950 (TGLA227) and 2,175 (ILSTS058).
Genotyping was performed by multiplex PCR amplification in a 15-µL reaction consisting of 4 µL of 10 ng/µL of DNA, 10 to 50 pmol of each primer, 200 µM of each dNTP, 1x PCR buffer (including 10 mM Tris-HCl [pH 9.0], 50 mM KCl, and 1.5 mM of MgCl2), and 0.2 to 0.4 units of Taq-polymerase (Qiagen, Valencia, CA). The 5'-end of each forward primer was labeled with one of the following fluorescent dyes: 6FAM, HEX, TET, PET, or NED (Applied Biosystems Inc., Foster City, CA). Amplification of fragments was performed in 96-well plates in a Techne thermocycler (InterMountain Scientific, Kaysville, UT) using PCR cycling conditions and annealing temperatures specific for each marker (52 to 61°C). Alleles were separated on an ABI 3730 Sequencer (Applied Biosystems, Inc.) at the Nevada Genomics Center, Reno.
Estimation of Probabilities of Exclusion
Probabilities of exclusion for the j-th microsatellite were estimated for each ranch and across ranches according to Jamieson (1994) and Jamieson and Taylor (1997) as
 | [1] |
where pi = the allele frequency of the i-th allele, and n = the number of alleles at the j-th microsatellite. The joint probability of exclusion for the 15 microsatellites was computed as
 | [2] |
All computations were carried out using software written in FORTRAN90 for this specific project.
Exclusion probabilities were also computed after dropping 1 to 5 microsatellites with the lowest probabilities of exclusion across ranches or for each individual ranch. This was done to establish a minimum number of markers required to achieve a probability of exclusion of at least 0.99 and to investigate if the microsatellites to be discarded will be the same across breeds and ranches. This strategy makes irrelevant which breeds were represented in the DNA paternity identification.
Benefit-Cost Analysis
The costs associated with a paternity identification program are determined by the genotyping effort:
 | [3] |
where m = the number of microsatellites used for paternity identification, g = the cost of typing one single microsatellite using multiplexing, Nb = the number of bulls at the ranch, 
= the proportion of cows pregnant in the herd in the year when genotyping is performed, and Nc = the number of cows. Therefore, Nc is the number of calves produced in the year in which paternity testing is carried out. The main benefits of establishing a paternity testing program specifically focused on herds maintained on open range (i.e., high, cold desert rather than pasture) result from identifying and culling of bulls siring progeny with low weaning weights. This can be evaluated as the genetic gain produced in the following year as a result of paternity identification and culling of poor performing bulls. The change in the mean in a given year, y, after culling bulls in year y – 1 can be evaluated using the equation for response to selection:
 | [4] |
where i = the selection intensity, h = the square root of the heritability (h2) of weaning weight, 
p = the phenotypic SD of weaning weights, and r = the accuracy of progeny testing with p progeny with value:
 | [5] |
Following Poutous and Vissac (1962), the discounted accumulated response at time horizon t is
 | [6] |
where d = the discount rate needed to account for inflation, and Ry is accumulated response at year y. The DAR measures the accumulated benefits in today’s current economic units by accounting for the inflation rate at the time the expression of the trait is attained. Benefits would begin in yr 1 if culling was initiated in yr 0. Since beef cattle are mated by natural service in our model, the maximum time horizon t is 3 yr. Otherwise, bulls would be mated to their own daughters.
Gomez-Raya and Klemetsdal (1999) proposed a benefit-cost analysis in marker assisted selection programs. Cost benefit analysis can be carried out by measuring benefits (increased performance in weaning weights) over cost (genotyping) as
 | [7] |
where 
is the market value of 1 kg of liveweight at calf weaning, and , Nb, GC, m, and g are as defined in Eq. 3. Note that the cost of genotyping is computed at yr 0, but the discounted accumulated responses are expressed in yr 1 to 3, where the benefits of culling bulls at yr 0 are obtained. Expression [7] assumes that remains constant over the years and gives the discount benefit in dollars per dollar invested in paternity identification.
For all benefit-cost calculations, a heritability and phenotypic standard deviation for weaning weight of 0.3 and 23.8 kg, respectively, was assumed. These values were derived by averaging the estimates from 3 crossbred populations located at the USDA Meat Animal Research Center, as reported by Dodenhoff et al. (1999). Three annual bull culling percentages, 10, 20, and 30%, were investigated. An inflation rate of 0.05 was used in all scenarios. An approximate current market situation was evaluated with $2.75 per kg ($1.25 per lb) of liveweight at weaning. If we assume the cost of genotyping is $1.00 per microsatellite, then the corresponding value for /g is 2.75. This cost of genotyping represents the cost of an optimized paternity identification program.
Breakpoints of /g for a DNA Paternity Program to be Profitable
The ratio of a unit of benefit (market value of 1 kg of liveweight) over a unit of cost ($ per microsatellite genotyped) is /g. In this section, we derive the breakpoint of /g at which a DNA paternity testing program is profitable. Let p be the actual average number of progeny per bull in the herd with value Nc/Nb. Substituting Nb = Nc /p into Eq. 7 and rearranging gives:
 | [8] |
This equation does not depend on the number of calves or bulls, but on the ratio of the number of calves to the number of bulls, . The first term in expression [8] is the ratio of the price of 1 kg of weaning weight to the cost of genotyping 1 microsatellite. This ratio can fluctuate with market conditions and improved DNA genotyping technologies. The second term in expression [8] depends on the ratio of the number of bulls to the number of calves, number of microsatellites, and the culling policy applied. Equation [8] can be solved for 
after making BC=1, to yield
 | [9] |
Expression [9] represents the breakpoint for the program to be profitable, because for BC=1, cost and benefit are equal. We evaluated the breakpoints in expression [9] that would make the paternity testing program profitable for 15 microsatellites genotyped, while varying p and bull culling percentages (i.e., 10, 20, and 30%).
Benefit-Cost Analysis with Incomplete DNA Paternity Identification
A large part of the cost involved in a DNA paternity testing program is in the genotyping of microsatellites. We investigated the benefit-cost of a program that uses a reduced number of DNA-markers and, consequently, has probabilities of exclusion less than 0.99. Assume that a set of m DNA-markers with joint probability of exclusion, PE, is used to reject alleged sires. Following Weir (1990), the posterior probability of being the true bull for any nonexcluded bull in the herd is
 | [10] |
where 
o is the prior probability of the alleged sire being the actual sire, and Pj is the probability of exclusion of the jth marker assuming typing of calves and bulls but not dams. We will assume that all nonexcluded bulls have the same prior probability of being the sire, and therefore, o=1/Nb.
On the other hand, the effect of incomplete DNA paternity identification on the estimation of a bull’s breeding value based on average progeny performance is that (1 – 1)p calves will not be offspring of the bull. It is shown in the Appendix that the accuracy of a progeny test based on a mix of progeny and nonprogeny from a bull is
 | [11] |
Equation [11] is simply the accuracy of evaluation of a progeny test based on the expected average number of progeny of the bull. Accuracy of evaluation depends ultimately on the number of DNA markers and their probabilities of exclusion as given in equation [10].
The benefit-cost equation with incomplete DNA paternity identification becomes
 | [12] |
This equation depends on the posterior probability of the bull being the actual father, 1, which is computed using expression [10] for a given set of DNA-markers.
We investigated the benefit-cost ratio of a program with incomplete DNA paternity identification with 20 bulls, 20 progeny per bull, and a ratio of benefits over cost 
of 2.75.
Probabilities of exclusion will generally be different for different beef cattle herds and DNA marker panels. We have computed the benefit-cost ratio of a program for 10 or 12 DNA markers with varying probabilities of exclusion (0.80 to 0.99), and varying number of bulls (10, 20, 40, 100). The yearly bull culling rate was 20% and 
. Thus, benefit-cost ratio can be approximated for any beef cattle operation and DNA marker set.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
. The number of alleles ranged from 10 to 19 with an average heterozygosity of 0.855. The probabilities of exclusion for each of the 15 microsatellites and 8 ranches are given in Table 3. Most of the probabilities of exclusion of individual DNA markers are consistently high on individual ranches. The joint probability of exclusion of the 15 microsatellites across ranches was greater than 0.99, and therefore, the 15 microsatellites can be used for paternity identification on Nevada beef cattle ranches.
|
|
show that dropping 1 or 2 microsatellites from the analysis does not have a big impact on the joint probability of exclusion. Using less than 13 microsatellites would result in probabilities of exclusion lower than 0.99, which means that paternity of some calves might not be accurately assessed.
|
On the other hand, Figure 1 depicts the breakpoints for market ratio prices of /g for a paternity program to be profitable for varying p and bull culling rates. DNA paternity testing was performed with 15 microsatellites. The bull culling rate is the single most important factor affecting the profitability of a DNA paternity testing program. The breakpoint for the program to be profitable at a culling rate of 30% is around 
, which means that benefits would accrue for market values above $1.10 per kg of liveweight at weaning, below $1.00 per microsatellite genotyped, or both.
|
shows the benefit-cost ratio of incomplete DNA paternity identification using DNA markers for 3 culling policies. In this scheme, a lower number of markers are used than necessary to achieve a joint probability of exclusion of 0.99, and account is taken of the reduction in accuracy of evaluation for the progeny test. The use of a low number of microsatellites results in greater benefit-cost ratios.
|
. A general situation is given in Table 6 that is applicable to any beef cattle operation irrespective of the genetic background or market conditions. A total of either 10 or 12 microsatellites are typed with probabilities of exclusion depending on the number of alleles and their frequencies for those markers in a given herd. The ratio is 1, and culling rate of bulls is 20%. The number of bulls is 10, 20, 40, or 100. For example, if the joint probability of exclusion for 10 markers is 0.98. Then, the benefit cost ratio is 0.88 for 10 bulls assuming according to Table 6. If actual market conditions are 
, then the benefit cost ratio for that beef cattle operation is 2.42 (0.88 x 2.75).
|
also shows 2 important features of the economic analysis of DNA paternity testing programs in beef cattle. The first is that a lower number of markers consistently yields greater benefit-cost ratios (approximately 20%). The second and more important feature is that the number of bulls in the herd has a dramatic impact on the benefit-cost ratio when probabilities of exclusion are below 0.95. In any circumstance, increasing the number of bulls results in reduced benefit-cost ratios. |
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
For the Nevada beef cattle ranches, the first question addressed in this paper is whether 15 highly polymorphic microsatellites could be used for DNA paternity identification, and whether the same markers could be effectively used across multiple ranching operations. Commercial beef cattle operations vary in breed composition, and therefore, variation in allelic frequencies of the DNA markers is expected. The 15 microsatellites used in this study had a large number of alleles and high heterozygosity (ranged from 0.700 to 0.923), which support their use for paternity analysis. Even the exclusion of 2 of the 15 microsatellites would result in reliable DNA paternity identification with a marginal loss in the joint probability of exclusion.
The International Society for Animal Genetics (ISAG) has recommended a set of microsatellite loci for routine use in bovine parentage testing and identification: ETH225, ETH10, TGLA227, ETH3, TGLA122, INRA023, BM2113, TGLA53, CSSM66, TGLA126, BM1824, and SPS115. Excluding marker CSSM66, the heterozygosity of this panel ranged from 0.588 to 0.862 using 26 commingled beef bulls and their calves from the Nebraska Reference Herd-1. This herd was predominantly composed of Red Angus cattle (Sherman et al., 2004). The panel of microsatellites used in our study had greater levels of heterozygosity, which might be attributed to the high number of different breeds and breed combinations in these herds.
A DNA paternity testing program would be profitable when the benefits of an increase in overall calf weaning weight override the cost of genotyping. The benefit-cost ratio was evaluated under a variety of economic circumstances, including an approximation of today’s market prices of calf liveweight and cost of genotyping. A DNA paternity identification program is clearly profitable when culling percentages of unproductive bulls are 20% or greater.
A program using a low number of microsatellites is cheaper (lower genotyping cost), but may not lead to full identification of paternities, and, consequently, might reduce accuracy of evaluation and response to selection (less benefit). However, we showed that incomplete DNA paternity identification may result in larger benefit-cost ratios (around 20% increase for 10 vs. 12 microsatellites). Benefit-cost ratio in DNA paternity identification programs in herds with a large number of bulls (and cows) would be reduced compared with herds with a small number of bulls. This is expected because more genotyping effort is needed to reject alleged parents with an increasing number of bulls. Beef cattle operations with a large number of bulls may subdivide the bulls and cows into small groups for breeding. In this way, benefit-cost ratio of DNA paternity testing would increase. Although benefit-cost ratio is greater in programs with incomplete paternity identification, total return of the program is greater when using a large number of microsatellites such that paternity is fully identified.
A wide variety of business models underlie beef cattle operations in Western North America. For example, some ranches eliminate a few bulls from the herd every year, whereas others replace all bulls every 4 to 5 yr. More research is needed to optimize DNA paternity programs in a variety of scenarios regarding proportion of bulls replaced in the herd each year. In addition, other economic aspects could be considered, such as the costs associated with culling bulls, the benefits of culling infertile bulls, and the possibility of using the resulting DNA information for traceability and national identification systems. Some of these factors are difficult to evaluate because they may vary across beef cattle operations.
Paternity identification using DNA markers opens new opportunities for breeding beef cattle on rangelands. For example, bull EPD for liveweight could be estimated based on progeny testing of their calves. Trade or sale between ranchers could be based on bull EPD.
The DNA paternity identification is feasible in beef cattle operations using 12 highly polymorphic DNA-markers. The DNA paternity identification and progeny testing may be profitable for open-range beef cattle operations with bull culling rates of 20% or greater. The benefit-cost ratio might be greater in a program with incomplete DNA paternity identification than in a program with full paternity identification. The benefit-cost ratios are greater for herds with a low number of bulls.
|
APPENDIX |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
The accuracy of evaluation is defined as the correlation between true (A) and estimated (Â) breeding values and is given by
 | [A1] |
For a progeny test based on performance of progeny (P1, P2, ...P3) and nonprogeny (P'1, P'2,...P'3) of the bull, the covariance between true and estimated breeding value is
 |
Assuming that no relationships exist between the bull and nonprogeny in his progeny test (Cov(A, P'1) = Cov(A, P'2) =...= 0), the covariance becomes
 | [A2] |
The variance of the estimated breeding value based on a progeny test with incomplete DNA paternity identification is
 | [A3] |
Substituting [A2] and [A3] into [A1] gives
 |
|
Footnotes |
|---|
2 Corresponding author: lgomezraya{at}cabnr.unr.edu
Received for publication January 27, 2007. Accepted for publication September 13, 2007.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXLITERATURE CITED |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
