MULTIPLE OVULATION AND EMBRYO MANIPULATION IN THE IMPROVEMENT OF BEEF CATTLE: RELATIVE THEORETICAL RATES OF GENETIC CHANGE 1 ABSTRACT

Transcription

1 MULTIPLE OVULATION AND EMBRYO MANIPULATION IN THE IMPROVEMENT OF BEEF CATTLE: RELATIVE THEORETICAL RATES OF GENETIC CHANGE 1 W. W. Gearheart, C. Smith and G. Teepker University of Guelph 2, Ontario, Canada N1G 2W1 ABSTRACT Theoretical rates of annual genetic responses to selection in beef cattle were compared for conventional and multiple ovulation and embryo transfer (MOET) breeding schemes. Several combinations of replacement policy, mating ratio and type of selection were considered for both schemes with low, medium and high heritabilities. For MOET, four rates of embryo transfers per donor were used to represent low to moderate MOET levels. The results indicated that annual genetic responses to selection could be up to 1.3, 1.6 and 1.8 times as great for MOET compared with conventional breeding for traits of low, medium and high heritability, respectively; however, the annual inbreeding rates also were high for the MOET schemes considered. Embryo splitting, or cloning, was shown to increase accuracy of selection by 8 to 35% through the production of identical genotypes. The use of MOET in conjunction with embryo splitting in elite nucleus units could substantially increase genetic improvement for traits with low, medium and high heritabilities in beef cattle populations. (Key Words: Beef Cattle, Embryo Transfer, Genetic Analysis, Selection Responses.) J. Anita. Sci : Introduction The potential for using multiple ovulation and embryo transfer (MOET) schemes to increase rates of genetic response in dairy cattle (Nicholas and Smith, 1983; Christensen and Liboruissen, 1986; Colleau, 1986; Woolliams and Smith, 1988), sheep (Smith, 1986) and swine (Smith, 1984) has been investigated. In general, MOET has the greatest potential for species with naturally low reproductive rates, such as cattle and sheep. Land and Hill (1975) considered the possible use of MOET schemes in beef cattle and found that theoretical annual selection intensities were greater for MOET schemes compared with conventional schemes. 1This research was supported by the Ontario Ministry of Agriculture and Food, by the Natural Sciences and Engineen~ng Research Council of Canada and by Semex Canada. /-Centre for Genetic Improvement of Livestock, Dept. of Anim. and Poult. Sci. Received December 23, Accepted March 10, Church and Shea (1977) reported that theoretical responses to selection could be doubled through the use of MOET breeding schemes. Whereas the work of Land and Hill (1975) and Church and Shea (1977) showed the potential for nearly doubling genetic response using MOET schemes, the comparisons were made for a single heritability and unspecified selection type. Little work has been done concerning MOET use in beef cattle since Many advances in MOET methodology have been made over the past several years, and embryo splitting and cloning now are possible. In light of these and other improvements in MOET and related technology, Land and Hill's (1975) work has been extended here to compare theoretical annual genetic responses for three types of selection and levels of heritability (h 2) for conventional and MOET schemes. In addition, use of embryo splitting, or cloning, for beef cattle is considered. Finally, the use of elite nucleus units for the improvement of commercial beef herds is proposed. 2863

2 2864 GEARI-IEART ET AL. TABLE 1. PARAMETERS AND SYMBOLS USED FOR CONVENTIONAL AND MULTIPLE OVULATION AND EMBRYO TRANSFER (MOET) BREEDING SCHEMES Scheme Parameter Symbol Conventional MOET Embryo transfers per donor k 4, 8, 12, 16 Embryo (progeny) sauvival to selection s.8.5 Generation interval a L L m = 2.5, 4.5 L m = 2, 4 Lf = 2.9, 3.4, 3.8, 4.2 Lf = 2 Heritability h 2.1,.3,.5.1,.3,.5 Inbreeding rate per year b F F = [(1/Nn0 + (1/Nf)] x 1/(8 x L) Mating ratio x 10, 20, 50, 100 2, 4, 6, 8 Number of progeny C 160 ks Donors per year D D = C/ks Proportion selected c P Pm = 1/(sx) Pm = 2/(ksx) Pf = 2/s(1 + E.9Y) Pf = 2/(ks) Selection intensity i Splits per embryo v 2, 3, or 4 asubscripts m and f refer to males and females, respectively. t'n refers to the number of animals. CSuperscript y refers to the number of year a female is kept in the herd and.9 is the proportion of surviving females per year. Materials and Methods Parameters and symbols used throughout this paper are given in Table 1 for reference. The current breeding structure of the beef population of North America consists of a large commercial segment (commercial herds, CH) and a smaller purebred segment (breeding herds, BH) that provides a large percentage of males used in commercial herds, either by AI or by natural service (Willham, 1988). Proposed elite nucleus units (NU) would improve commercial herds either directly through AI or natural service males or indirectly through current BH. Conventional Breeding Schemes. Throughout this paper, conventional schemes will refer to beef cattle reproduction without MOET techniques. Beyond this restriction, the conventional schemes outlined could be generalizations of any number of current breeding herds in North America. A constraint on the number of progeny that can be produced per year was assumed and a constant number of females was maintained at 200 per year. The probability (s) of an offspring surviving to time of selection was.8, resulting in 200 x.8 progeny per year (C) available for selection. The probability (d) of a calving female surviving to the following year was.9. Females were used from 2 to 4, 2 to 5, 2 to 6 and 2 to 7 yr and mating ratios (x) of 1:10, 1: 20, 1:50 and 1:100 were considered. Males had generation intervals of 2.5 or 4.5 yr and 20 progeny were tested per male (n). MOET Breeding Schemes. For MOET schemes, mating ratios were reduced to 1:2, 1: 4, 1:6 and 1:8 to reduce inbreeding. Four rates of embryo transfer per donor (k) were considered (4, 8, 12 and 16). For MOET breeding, s was taken to be.5 to allow for unsuccessful ET. The proportion of males (Pm) and females (pf) selected for breeding per year were 2/ksx and 2/ks, respectively. To have similar herd sizes for comparison with conventional schemes, the number of donors per year (D) was assumed to be C/(ks). Males were 2 or 4 yr old as parents and females were 2 yr old as parents. For MOET progeny testing, n again was 20. Selection Types. The types of selection considered for comparison of annual genetic responses of conventional and MOET schemes are individual (I), family index (II) and family index followed by progeny testing (III). The family index contains information on an individual and its sibs, paternal half-sibs for the conventional systems and full and paternal half-sibs for MOET systems. Progeny testing selection schemes were based on a sequential culling program as outlined by Morris et al. (1980). Young males initially were selected based on family indices of performance records (selection type II), their progeny were tested and progeny-tested males were selected and used for mating. In all cases, the trait under selection was assumed to be expressed in both sexes with equal heritability and

3 MULTIPLE OVULATION/EMBRYO TRANSFER IN CATI'LE 2865 genetic correlations of expression between sexes were unity. For selection types I and II, annual response to selection (R) is estimated by: R = [((i m + if) x rig x h)/(l m + Lf)]x o (1) where i m and if are male and female selection intensities, Lm and Lf are male and female generation intervals and rig is selection accu- racy. With type III selection, allowance must be made for the sequential selection applied to males. The genetic merit of selected females (Gf) is if x rig x h x t~, but the genetic merit of males selected on the progeny test (Gm) can be estimated (Morris et al., 1980) as: Gm = qg x h x [iml + ira2 x q] x a (2) where q = ~[1 - kl + {n(e)2/(4(l+.25h 2) + nh2(e) - h4(1 - k2))}] (3) k 1 = 1 - imt(iml - Xml) (4) k2 = 1 - if(if - xf) (5) In (2) through [5] above, e is 1 - h 2, rig is accuracy of selection type II, n refers to the number of progeny tested per male, x is the truncation point at first selection, iml and im2 are male selection intensities of the first and second selection, respectively, and if is female selection intensity. Annual response for selection type III then can be estimated as: R = [(Gm + Gf)/(Lm + Lf)]. (6) For comparison of conventional and MOET schemes, three heritability levels were considered,.1,.3 and.5. For all schemes, progeny testing for type III selections was assumed to occur outside the herd to remove any constraints due to herd testing facilities. It was assumed that some selection had been applied to females used for male progeny testing. A sample calculation for a conventional type III selection scheme is given in appendix A. Inbreeding. Annual inbreeding rate (AF) was calculated for all schemes as: AF = [(1/Nm) + (1/Nf)/(8 x L)] (7) where N m and Nf are the number of breeding males and females in the herd each year and L is mean generation interval. For these MOET schemes, Nm = D/x and Nf = D. Formula (7) likely underestimates AF due to a tendency to select more individuals from the best sibships (Hill, 1976; Woolliams and Smith, 1988). Embryo Splitting. To determine the benefits to selection from embryo splitting, or cloning, increases in accuracy of type II selections were measured when information on two or four splits of an embryo was included in the index and compared to an index having information on a single embryo for MOET schemes. Heritabilities of.1,.3 and.5 were used with intraclass correlations (ICC), or the correlation between splits in this case, equal to or double h 2.. Resul~ Annual Response to Selection. Predicted annual rates of response to individual (I) and family index (II) selection and II followed by a progeny test (III) selection for conventional breeding are shown in Table 2. For selection types I and II, the greatest annual progress is expected to be at the highest mating ratio of 1: 100, with females calving for 2 to 5 or 2 to 6 yr. For these cases, the expected responses to selection type I were.052,.154 and.257 phenotypic SD units per year for h 2 of.1,.3 and.5, respectively. Responses to selection ~2PeolI were.082, 182 and.272 SD units for f.1,.3 and.5", respectively. The highest response to selection type III was at a mating ratio of 1:100, cow ages of 2 to 3 and a h 2 of.5. As cow ages increased, annual response from selection type III decreased due to the increased generation intervals. Except for cow ages 2 to 7 at a mating ratio of 1:10, response to selection type III always was greater than the responses to selection types I and II for h 2 of.1. At h 2 of.3 and.5, increasing female generation intervals resulted in larger re-

4 2866 GEARHEART ET AL. TABLE 2. PREDICTED ANNUAL RESPONSE a FOR CONVENTIONAL BREEDING SCHEMES AT THREE HERITABILITIES (h 2) AND FOUR COW RELACEMENT POLICIES (COW AGES) Mating ratio and selection b Cow ages, yr h 2 I 1I Ill I II III 1 II III I I1 III aphenotypic sd units x 103. bselection types I, II, and III refer to individual index, half-sib family index and II followed by progeny testing, respectively; male generation interval was 2.5 yr for types I and II, 4.5 yr for type III. sponses for type II vs type III selection for several cases. Given in Table 3 are the predicted annual genetic responses for MOET breeding schemes using the selection types described above, with the addition of full-sib information to the family indices. Even at moderate MOET rates (ks = 4) with a mating ratio of 1:8, annual response to each selection type was greater than the best possible under conventional schemes at all levels of h 2. As expected, highest responses shown for the MOET TABLE 3. PREDICTED ANNUAL RESPONSE a FOR MULTIPLE OVULATION AND EMBRYO TRANSFER (MOET) BREEDING SCHEMES AT THREE HERITAB1LITIES (h 2) AND FOUR EMBRYO TRANSFER (ks) RATES b Mating ratio and selection c ks h 2 1 II Ill 1 II III I II III I II III asd units x 103. bthe number of embryo transfers per donor (k) times the probability (s) of an embryo surviving to time of selection. CSelection types I, II, and III refer to individual index, half-sib and full-sib family index and II followed by progeny testing, respectively; generation intervals were 2 yr for males of type I and II and 4 yr for males of type III. All female generation intervals were 2 yr.

5 MULTIPLE OVULATION/EMBRYO TRANSFER IN CATFLE 2867 TABLE 4. PREDICTED ANNUAL RATE OF INBREEDING (AF) FOR CONVENTIONAL BREEDING SCHEMES ab Mating ratio and selection r Cow ages I,II 111 I,II 1II I,II III I,II III amales generation interval is 2.5 yr; number of females was 200. bvalues given are x 104. CAnnual inbreeding was the same for selection types I and II. schemes were at the parameter extremes, ks = 8 and a mating ratio of 1:8, where the MOET responses were from 1.2 to 1.8 times those obtained by conventional schemes. Inbreeding. The annual rate of inbreeding (AF) quickly becomes large for MOET schemes of limited size. Shown in Tables 4 and 5 are the estimated AF levels for conventional and MOET schemes of 160 calves/yr. For conventional breeding using a mating ratio of 1:20, AF ranges from.0078 to.0095 for selection type I and II and from.0067 to.0080 for selection type III. However, at the lowest ET rate considered and a mating ratio of 1:2 for MOET, the AF was.0059 and quickly increased to.0199 for types I and II as the number of ET increased from 4 to 16. At the highest ET rate considered and a mating ratio of 1:8, AF was.0598 for type I and II and.0398 for type III selection. Embryo Splitting. Splitting of beef embryos would lead to increased selection accuracies due to having information on identical genotypes. Given in Table 6 are proportional increases in the accuracy of family index selection (type II) when information in in- cluded on two, three, and four splits of an embryo. For these MOET selection indices, adding infonnation on splits resulted in the greatest increase in accuracies at the lowest h 2 and intraclass correlation levels. Because h 2 sets the lower bound for ICC, the greatest increase in accuracy would be expected when ICC = h 2, with small increases in accuracy expected as ICC approaches 1. For the schemes considered, for example, having two or four animals of each genotype increased the accuracy of family index selection by as much as 15 and 28% of h 2 of.1 but only 8 to 15% at h 2 of.5. Discussion Morris et al. (1980) found that a policy of selecting males first on performance then on a progeny test to be 26 to 38% better than individual selection in terms of annual genetic progress for a heritability of.3 and progeny testing outside the herd. For the conventional schemes outlined here, selection on performance followed by progeny-testing males (type III) resulted in annual responses as much ks TABLE 5. PREDICTED ANNUAL RATE OF INBREEDING (AF) FOR MOET BREEDING SCHEMES ab Mating ratio and selection c I,lI lit I,II III I,II III l,ii III amales generation interval is 2 yr; number of donors (D) -- C/ks, where k is the number of embryo transfer per donor and s is the probability that an embryo will survive to time of selection (s =.5). bvalues given are 104. r inbreeding was the same for selection types I and II.

6 2868 GEARHEART ET AL. as 54% greater than individual selection (type I) at the highest mating ratio of 1:100. Selection type III was as much as 28% better than selection type II at a mating ratio of 1: 100, although a mating ratio of 1:100 may be unreasonable and was used as an extreme case. At a mating ratio of 1:50 and h 2 of.3, type III selection was as much as 22 and 9% better than type I and type II selection, respectively. Although some of the technologies discussed are still difficult and expensive, genetic improvement in beef cattle would be possible through the establishment of an elite NU that would be maintained with MOET breeding schemes. Nicholas (1979) first proposed the use of a MOET nucleus unit scheme for dairy cattle consisting of elite males and females chosen from the population. For purposes of improving beef cattle, elite NU could be established by selecting genetically proven elite males and females from the breed population based on individual and pedigree information. In practice, NU would be open to other elite stock. In this discussion, the units are, for simplicity, assumed to be maintained TABLE 6. PERCENTAGE INCREASE IN SELECTION ACCURACIES WHEN EMBRYO SPLIT INFORMATION IS INCLUDED IN MULTIPLE OVULATION AND EMBRYO TRANSFER (MOET) SELECTION INDICES a as closed herds using MOET breeding schemes. Shown in Figure 1 are possible relationships among the proposed NU and existing herds. Herds are arranged in descending order of genetic merit, with the NU being superior due to initial genetic lift and MOET breeding. Current BH would be a level below this, with CH at the lowest genetic level. Current BH and CH could be improved through the use of NU males as A1 sires and the possible future transfer of genetic material by ET. Establishment of NU could result in a genetic lift equivalent to 3.4 yr of annual response, the amount of lift being dependent on selection intensities applied to the males and females initially chosen and on heritabilities (Nicholas and Smith, 1983). Operation of the NU at even moderate MOET rates in combination with embryo splitting would result in a high rate of genetic change. The BH and CH herds would be maintained with conventional breeding schemes. Males for BH herds would be selected on half-sib family indices or progeny tests from the NU. The CH herds would buy selected males from NU or BH herds. Both BH and CH herds should use NU males as AI or natural service sires. However, depending on who actually estab- Animals per identical Mating ratio h 2 ICC genotype aaccuracies compared are for selection type II at ks = 4; :Z 2 the intraclass correlation (ICC) was either h or 2 h for ease of calculation. High - MOET breeding - Selections A, H and C All AI NS [ NS ET i ET CH - Conventional breeding - NU and/'or BH males NS - Conventional breeding - Selections A and C Figure 1. Possible relationships among an elite nucleus unit (NU), a current breeding herd (BH) and a commercial herd (CH). AI refers to artificial insemination and NS refers to natural-service sires. Selections are based on male indices (A), female family indices (B) or male performance (A) followed by progeny testing of males (C). AI 1

7 MULTIPLE OVULATION/EMBRYO TRANSFER IN CATrLE 2869 lishes and maintains the NU, NU AI and natural service sires could be used exclusively for improvement of BH. This BH improvement eventually would be disseminated throughout the CH, but genetic lag for CH improvement would be longer if only BH herds were used. In addition to the use of NU sires, the possibility of transferring NU embryos exists for future improvements in BH and CH herds. An important aspect to consider in establishing NU concerns the responsibility for establishing and maintaining the units. One possibility would be for current breeders or groups of breeders to found and run such units; this would lead to competition between the unit breeders and other breeders for the commercial herd market. Another possibility would be for an interest outside the beef cattle industry, government or private, to fund the establishment and maintenance of nucleus herds, which then would be in competition with all current breeders and possibly would lead to further benefits for commercial operators. Smith (1988a) has suggested the possibility of maintaining several elite units, reducing the unit sizes and costs per unit and removing the risk of stock and resource concentration into a single unit; however, the reduction in unit sizes would lead to increased inbreeding in each unit and other problems associated with small population numbers (Nicholas, 1980). One of many possible breeding schemes to assure the continued high rate of improvement in NU, and therefore in BH and CH, would be to use performance-progeny-tested males and to use MOET on young females chosen from references sites (referenced sires used on many herds). This system would yield an additional amount of genetic gain through the turnover of young females. Even though there are definite advantages to the establishment of elite NU, such as a genetic lift, faster rates of genetic change and possible use of new or expensive technologies, some potential disadvantages exist (Smith, 1988b). These included disease risks, risk from concentrating all stock and resources in one unit and possible genotype x environment interactions. Inbreeding can be a problem with small MOET schemes. However, moderate annual inbreeding rates are tolerable. For example, the calculated yearly inbreeding rate (AF) for a NU at a moderate MOET rate of ks = 4 and a mating ratio of 1:2, assuming the number of females is 200, would be about.004. This level of inbreeding may be tolerable, because the annual response for h 2 of.3 would be about.143 for selection type III. If the number of females were 400, however, AF is reduced to approximately.002, or only.06 after 30 yr of selection. Although the annual inbreeding obviously accumulated faster for the MOET schemes outlined here, it can be difficult to interpret AF alone. Dr. John Gibson (personal communication) has suggested that the ratio of AF per unit of response may be a more meaningful expression for interpreting AF for a breeding scheme. Nicholas and Smith (1983) suggested selecting only one individual per full sibship as a means of moderating AF. In addition, increasing the number of females and keeping the mating ratios low will reduce AF. Advantages from MOET breeding are largely due to higher selection intensities that can be applied to females with MOET and make short generation intervals possible (2 yr). Theoretical response rates estimated for those MOET schemes are as much as 1.8 times those possible for conventional schemes, or slightly less than double, previously reported (Land and Hill, 1975; Church and Shea, 1977) for a heritability of.5. However, for h 2 of. 1 and.3, MOET increased predicted annual responses only 1.3 and 1.6 times, respectively. Although some of these MOET rates (ks = 6 or 8) were moderate or even high compared with current standards, they well may be feasible in the future. Annual response rates achieved in practice generally are less than those shown by theory. Smith (1988a) discussed many of the reasons for this discrepancy, including selection disequilibrium (Bulmer, 1971), reduced selection differentials in small populations (Hill, 1977), inbreeding and genetic drift. Splitting embryos to produce twins or continued splitting to produce several clones of an embryo has been accomplished with some success in cattle (Donahue, 1986; Marx, 1988). As indicated in Table 6, including information on embryo splits could be beneficial to beef cattle breeding schemes through increased selection accuracies and therefore increased responses. In practice, however, potential advantages of embryo splitting are reduced because the survival rate of split embryos is low. For example, assuming a conceptionsurvival rate for split embryos of.5, only.25

8 2870 GEARHEART ET AL. of the embryos split would result in an identical pair for evaluation. Also, additional testing space would become necessary. Woolliams (1988) found that under a constraint of fixed testing resources, cloning would add to genetic progress only at low h 2 or at the expense of a reduction in the number of male families, which would increase inbreeding. Even if splitting and cloning techniques are not immediately implemented into current BH or CH, an elite NU could use these techniques to increase the dissemination rate to BH and CH. Improvement to be realized from embryo cloning and storage in beef cattle breeding would arise through selection of parents of the embryos to be cloned, selection of the best clones to be used and increased rates of annual genetic responses to selection (Smith, unpublished data.). Implkmt ons As research and technological developments provide new and improved methodologies for the various procedures considered, multiple ovulation and embryo transfer and embryo splitting may become commonplace and more economically feasible. Implementation of these technologies in established elite nucleus units would lead to substantial improvement in performance for reproduction, growth and carcass traits. Further investigation, using computer simulations, is necessary to determine realistic response rates and to determine optimal multiple ovulation and embryo transfer policies for achieving improved selection responses for traits of economic importance in beef cattle populations. Literature Cited Bulmer, M. G The effect of selection on genetic variability. Am. Nat. 105:201. Christensen, L. G. and T. Liboriussen Embryo transfer in the genetic improvement of dairy cattle. In: C. Smith, J.W.B. King and J. C. McKay (Ed.) Exploiting New Technologies in Animal Breeding: Genetic Developments. pp Oxford Univ. Press, Oxford. Church, R. B. and B. F. Shea The role of embryo transfer in cattle improvement programrnes. Can. J. Anim. Sci. 57:33. Colleau, J. J Genetic improvement by embryo transfer within an open selection nucleus in dairy cattle, ha: O. E. Dickerson and R. K. Johnson (Ed.) 3rd World Cong. on Genet. Appl. to Livest. Prod. pp Lincoln, NE. Donahue, S. F A technique for bisection of embryos to produce identical twins. In: J. W. Evans, A. Hollander and C. M. Wilson (Ed.) Genetic Engineering of Animals; An Agricultural Perspective. pp Plenum Press, New York. Hill, W. G Order statistics of correlated variables and implications in selection programmes. Biometrics. 32: 889. Hill, W. G Order statistics of correlated variables and implications in genetic selection programmes. II. Response to selection. Biometrics 33:703. Land, R. B and W. G. Hill The possible use of superovulation and embryo transfer in cattle to increase response to selection. Anim. Prod. 21:1. Marx, J. L Cloning sheep and cattle embryos. Science 239:463. Morris, C. A., L. P. Jones and I. R. Hopkins Relative efficiency of individual selection and reference sire progeny test schemes for beef production. Aust. J. Agric. Res. 31:601. Nicholas, F. W The genetic implications of multiple ovulation and embryo transfer in small dairy herds. Proc. Conf. Eur. Assoc. Anita. Prod., Harrogate, UK. Nicholas, F. W Size of population required for artificial selection. Genet. Res. (Camb.) 35:85. Nicholas, F. W. and C. Smith Increased rates of genetic change in dairy cattle by embryo transfer and splitting. Anim. Prod. 36:341. Smith, C Biotectmology in animal breeding programmes. Proc. 2nd World Cong. on Sheep and Beef Cattle Breeding. Kopie-rite, Pretoria. Smith, C Use of embryo transfer in genetic improvement of sheep. Anim. Prod. 42:81. Smith, C. 1988a. Checking rates of genetic response with new reproductive techniques. Proc. 3rd World Cong. on Sheep and Beef Cattle Breeding, Paris. Smith, C. 1988b. Applications of embryo transfer in animal breeding. Theriogenology 29:203. Willham, R. L Selection objectives and programs applied to beef breeds in order to improve efficiency: North American example. Proc. 3rd World Cong. on Sheep and Beef Cattle Breeding, Paris. Woolliams, J. A The value of cloning in MOET nucleus breeding schemes for dairy cattle. Anim. Prod. (ha press). Woolliams, J. A. and C. Smith The value of indicator traits in the genetic improvement of dairy cattle. Anim. Prod. 46:333. Appendix A Calculation of Annual Response to Selection Type III for a Conventional Breeding Scheme. An example of the calculation of the annual response to type III selection (R) is given using formulas (2) through (6) above. For this example, females were used for 2 to 4 yr (y = the number of breeding seasons) with the probability (d) of a calving female surviving to the following year of.9. The probability (s) of an offspring surviving to selection was taken to be.8. The proportion of females selected (Pf) was determined from the number of replacement females (Rf) needed and the

9 MULTIPLE OVULATION/EMBRYO TRANSFER IN CAITLE 2871 number of females available where and Pf = Rf/Af, (Af) as: (A1) Rf = Nf/(1 + Edn), (A2) Af = (Nf/2) x s. (A3) In (A2) and (A3) above, Nf is the number of females in the herd. Then Pf = 2/( x x.9) =.92, and the female selection intensity (if) is.16. Finally, predicted female genetic merit (Gf) is: Gf =.16 x.55 x.55 x a =.048 6, for h 2 of.3. The number of males necessary is 10 with 200 females at a mating ratio of 1:10, because males were used for 2 yr, and the number of males available for selection is 1/2(200 x.8) = 80. The optimal selection policy, or the greatest total selection intensity (iml + im2), was determined by simulating all possible combinations of males chosen at each selection. The best policy was selecting 30/80 males initially and 10/30 males at the second selection, resulting in intensities of iml = 1.01 and ira2 = With n = 20 and h =.3, using (3) through (5) above, q = With an accuracy of selection for type II in this case of.60, the male genetic merit (Gin) using (2) is Finally, the female generation interval (Lf) can be found as: Lf = (.8 x x.9 x x.9 x 4)/( x x.9) = 2.93 and the male generation interval (Lm) is 4.5 yr, Then annual response to type III selection for this example is given by (6) as Annual response to selection type III with MOET schemes can be calculated using the same procedure as outlined above.