- Email: [email protected]
Breeding Strategies for Genetic Improvement of Dairy Cattle in Zimbabwe. 1. Genetic Evaluation
Breeding Strategies for Genetic Improvement of Dairy Cattle in Zimbabwe. 1. Genetic Evaluation NTOMBIZAKHE MPOFU,I CHARLES SMITH, and EDWARD B. BURNSIDE Centre for Genetic improvement of Livestock Department of Animal and Poultry Science University of Guelph Guelph, ON, Canada N1G 2W1 ABSTRACT
Abbreviation key: CEI strategy = untested young bull team produced from continual embryo importation; CSI = continual semen importation for 100% of the population; CSI30 = continual semen importation for 30% of the population; CSISO = continual semen importation for 50% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; MOET = multiple ovulation and embryo transfer; NMOET = scheme using MOET nucleus herd (closed); PT1 = closed progenytesting scheme; FT2 = progeny testing as PT1 but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as FT1 but with imports for sire of sons; rg = genetic correlation of performance between countries.
Strategies of selection within local populations and of importation of semen and embryos from countries with higher genetic merit are evaluated for genetic improvement of milk yield in Zimbabwe. Local programs are progeny testing in a closed population, progeny testing combined with semen importation to sire 30% of cows, progeny testing with foreign sires as sires of bulls, and a closed nucleus selection scheme using embryo transfer started from elite imported stock. Strategies based on continuous importation are imported semen for 30, 50, or 100% of cows and semen from elite foreign bulls used on local elite cows and bulls from elite imported embryos. Gene flow methods were used to estimate the genetic means for the strategies. In predicted genetic response at 25 yr, continuous importation of all semen from the superior exporting country ranks first; second is the importation of elite embryos to provide bulls for AI; and third is the closed nucleus scheme using embryo transfer. Schemes involving importation of semen to breed bulls with local progeny testing are intermediate, and the lowest responses come from semen imported to breed cows and from the closed local progeny-testing scheme. Ranking of the strategies is influenced by the genetic correlation of performance and by the initial genetic difference between Zimbabwe and the exporting country. (Key words: local selection, imports, breeding strategies)
INTRODUCTION
Received July 24, 1992. Accepted November 9, 1992. 'Current address: Department of Animal Science. University of Zimbabwe, Harare, Zimbabwe. 1993 J Dairy Sci 76:1163-1172
The two main methods of genetic improvement are selection and migration. Selection techniques, such as progeny testing, used in most developed countries should be applicable to Zimbabwe because genetic parameters for milk yield are similar (13) and herd recording is increasing. The rates of progress made through progeny testing, although low, are cumulative, thus resulting in improved milk yield over time. Given the costs of such selection programs, Smith (19) suggested that developing countries adopt breeding strategies based on importation or on nucleus breeding schemes using multiple ovulation and embryo transfer (MOET). Imports are recommended when 1) exotic stocks are genetically better than domestic stock, 2) when selection objectives are similar, and 3) when genotype by environment interaction is not important. In Zimbabwe, temperate breeds are used for commercial dairy production. The breeding objective in Zimbabwe is to increase genetic merit for milk yield. The 1990 national average yield was 5830 kg per cow per 300-d lactation (23).
1163
1164
MPOFU ET AL.
The Zimbabwe Holstein population was estimated to be genetically 1.25 SD [phenotypic (1 SD = 540 kg)] below the US population, and the genotype by environment interaction between the two countries was estimated by a genetic correlation of .70 f .18 (14). Therefore, strategies involving importation of genetic material from the US or other temperate countries should be evaluated. The objective in this paper is to evaluate genetically breeding strategies for Zimbabwe. Four strategies based on domestic selection programs and five based on importation of foreign genetic material are considered. The effects of genotype by environment interaction and the initial difference in genetic means between Zimbabwe and the exporting population on the ranking of the strategies are also studied. A complete evaluation of strategies should include an economic evaluation, which is the subject of a companion paper (15). MATERIALS AND METHODS
The parameters assumed for the Zimbabwe dairy population are a population of 100,OOO cows with 40% milk recorded, and, of these, 80% are bred by AI (Zimbabwe Milk Record Office, Harare, Zimbabwe, 1984 to 1990, annual and internal reports). Heifers calve for the first time at 24 mo of age; calving interval is 12 mo. Cows may survive up to five lactations, and the distribution by lactation number is 30, 25, 20, 15, and 10% for lactations 1 to 5, respectively. For simplicity, imports are assumed to be from one country, the US. The exporting country is assumed to have a very large dairy population, to have a long-standing, closed progeny-testing scheme with an annual rate of progress of .16 standard deviation, and to have selected cattle that are genetically superior to Zimbabwe cattle. Assumptlons for the Proposed Breeding Strategies
Progeny Testing. Three progeny-testing strategies are studied. The first strategy (PT1) is a closed progeny-testing scheme. Thirty-five young bulls are tested per year on 40% of the recorded cows bred by AI. Each test bull has 40 daughters with completed lactations. The top 3% of test bulls are selected as sires of Journal of Dairy Science Vol. 76, No. 4. 1993
sons and the top 9% as sires of replacements, each bull producing 40,000 doses of semen. The generation intervals are 2 yr for the test bulls, 6 yr for sires of replacements, and 7 yr for sires of sons. Only cows in their third and fourth lactations (ages 4 and 5 yr) are selected as bull-dams, and these two parities contribute equally to the next generation. Eight bull-dams are selected for every test bull required. No selection is practiced for dams of cows. The generation interval for the cows is 4 yr. The second strategy (PT2) differs from PT1 in that 30% of the cow population is sired by foreign bulls. Thus, a smaller bull stud than for PT1 is required. Only 20% of the cows are bred to test bulls, and each test bull has 20 daughters with complete lactations. The third strategy (PT3) is like PT1, but the best available foreign bulls are used as sires of sons. These bulls are all 6 yr old. MOET Nucleus Scheme. A MOET nucleus scheme (NMOET) is set up to produce and to test bulls for use in the commercial cow population. The nucleus herd is founded by screening and selecting stock with the highest estimated breeding values from the exporting population. The initial genetic mean of the nucleus is therefore higher than that of the Zimbabwean base population. Thereafter, selection is within the nucleus. The size of the herd is 50 donor cows. Donor cows are superovulated, giving 8 embryos per flush. With a pregnancy rate of 65% (obtained for fresh embryos in Zimbabwe), a calf survival rate of 80%, and a sex ratio of 50%, the average number of surviving male (and female) offspring per flush is 2. The cows are flushed four times and are mated to a different bull at each flush. Approximately 4 bulls are required to service the base (commercial breeding) population. Because the MOET bulls are selected on sibling information, selection accuracy is lower, so the rate of use per bull is reduced, and twice as many bulls as required are selected for the base population. A restriction of 1 bull selected per full sibship and 1 per dam results in a selection pressure of 4 out of 50 for nucleus sires and 8 out of 50 for sires for the base population. The selection pressure for nucleus females is 50 out of 200. The generation interval is 51 mo. Selection of bulls is on their full and half-sib sisters’ first lactation records.
1165
GENETIC EVALUATION OF BREEDING STRATEGIES
Continual Semen Importation. With continual semen importation, there would be no bull testing program in Zimbabwe. Semen from above average proven bulls (as opposed to elite bulls) is imported to sire replacements. The proportion of cows sired by foreign bulls is varied to include 30 (CSDO), 50 (CSISO), and 100% (CSI). The imports are made every year with each new batch of foreign proven sire semen higher in genetic merit than the previous batch because of genetic trend. When less than 100% of the cows are bred to foreign bulls, the rest of the cows are bred to untested local bulls. Local bulls are used from 2 to 6 yr of age, and the age groups contribute equally to the next generation. Even when semen is imported mainly to sire cows, some of the local bulls are sired by foreign bulls, contributing 10% of the untested local bulls for CSI30 and 20% for the CSISO strategy. Untested Young Bull Teams from Local Elite Cows and Elite Foreign Bulls. The ELITE strategy is based on continued importation of semen from elite bulls to sire sons. The young bulls are used untested. To reduce the risk from untested bulls, the rate of usage per bull is reduced from the 40,000 AI for progeny-tested bulls to 4000 AI per bull, so 40 young bulls are required. Bull-dams are elite cows selected from the local population; MOET is used, making it possible to produce at least 1 bull per bull-dam. Embryo collection is conducted in individual herds; therefore, embryo yields are expected to be lower than for NMOET. Pregnancy rates should be similar to those for NMOET because in both cases fresh embryos are used. Semen collection starts at 15 mo, and bulls are used for 1 yr. The female calves from elite matings are kept in the herds in which they are born and can be selected as bull-dams. Continual Embtyo Importation. In the CEI strategy, highly selected bull-dams in the exporting country are mated to elite bulls to produce embryos for export. Young bulls from these embryos are used untested for regular service at 15 mo of age, and the heifers are sold to commercial herds. As for the ELITE strategy, 40 young bulls are produced for the CEI strategy. Because the embryos will be frozen, the success rate (32%) is lower.
Method of Evaluation
The effectiveness of the different breeding strategies is assessed by comparing predicted genetic progress possible by adoption of each strategy using deterministic simulation methods. Rendel and Robertson (17) predicted genetic change based on four selection pathways: bulls to breed bulls, bulls to breed cows, cows to breed bulls, and cows to breed cows. The estimate of annual genetic change (AGy) is
AG,, = CAGEL
[I1
where AG is the genetic superiority for each of the four pathways, and L is the generation interval for each of the four pathways. The equation gives asymptotic response rates for overlapping generations. Hill (9) presented a gene flow procedure using matrix formulation based on the Rendel and Robertson (17) formula. These procedures are used to evaluate the breeding strategies. The evaluation criterion is the weighted genetic mean for cows in their first to fifth lactations, weighting by the proportion of the age group. The strategies are evaluated over a 25-yr period. The gene flow model used is
M(t) = P[M(t - 1)
+ AG(t)l
[2]
where M(t) is a vector of average genetic merit of animals at time t, P is a gene transmission matrix, and M(t - 1) is a vector of average genetic merit of animals at time (t - l), and AG(t) is a vector of the genetic superiority of selected animals. Continuous selection in both sexes is assumed. The P matrix specifies the contribution of different age classes to the next generation. The P matrix is assumed to be fixed for each strategy; i.e., the contributions of the different age classes to the next generation are assumed to be the same across years. The elements of the P matrix are defined to reach assigned generation intervals for each strategy. The initial average genetic merit [M(t = O)] is set to 0 for the Zimbabwe cow population. The genetic difference between the Zimbabwe cow population and that of the importing country is assumed to be 1.25 standard deviations (675 kg of milk). This value is used for the initial genetic difference between the NMOET and the Zimbabwean base population. Imports Journal of Dairy Science Vol. 76, No. 4, 1993
1166
MFQFU ET AL.
are either semen from tested bulls or embryos sired by elite bulls. Cows are assumed to lag behind selected bulls by 3AG (2, 15), where AG is the assumed annual genetic response. For the exporting country, the lag is .48 standard deviations (.16 x 3). The mean for the selected foreign bulls is therefore equal to 1.73 standard deviations (1.25 + .48). Elite bulls are assumed to be the top 10%. Semen from these bulls is imported for strategies PT3 and ELITE, and embryos for CEI are sired by elite bulls. Semen for strategies PT2, CSDO, CSI50, and CSI is from bulls that are .21 standard deviation above the mean or the top 42% bulls (14). The importation of semen from bulls in the top 10, 20, and 30% is investigated for pT2, CSDO, CSI50, and CSI. The within-group selection differential of the selected animals at time t, AG(i,t), is given by
where rg in milk yield is between the importing and exporting countries. The effect of rg on the ranking of breeding strategies is investigated. The range of values used is .5, .7, .9, and 1.0. The rg between performance in the nucleus and the base population is assumed to be 1.0 for NMOET.
AG = r x I x ug,
RESULTS
[31
The rate of inbreeding is calculated using Wright's formula (7). The inbreeding rates are scaled up by a factor of 2 for NMOET following Ruane and Thompson (18). The reduction in the mean per year because of inbreeding depression is calculated assuming .27% depression in milk yield per percentage of increase in inbreeding (12). The genetic superiority for imports to Zimbabwe is calculated as AG = I x r x u, x rg
[41
The genetic responses are calculated over where I is the intensity of selection, r is the years up to yr 25, using a standard set of and ug is the genetic accuracy Of parameters (initial difference, 1.25 SD, rg = standard deviation for milk yield. Heritability .7), and refer to the Zimbabwe commercial for Yield was assumed to be genetic cow population. The responses are then corncorrelation (r,) to be 1.00 between successive pared with pT1, the closed Zimbabwe lactations, and repeatability of yield over SUC- progeny-testing scheme. varicessive lactations to be e40. The results for the local selection strategies ance is standardized to 1.00. The accuracy of are in Figure 1. The response is highest with evaluation (r) for both bulls and COWS is calcu- NMoET: 66% higher than pT1 at yr 25, lated using selection index theory for bulls largely because of the initial genetic lift using daughter information and for cows using recouped by NMOET for the long-term genetic information from the she, the dam, and the response rates are similar. with p ~ 2in, which cow*s Own lactation For the "loET 30% of cows are bred by foreign bulls, an scheme, the accuracy of evaluation for bulls is early (but smaller) response is also obtained, calculated using sibling information and for but the long-term response rate is higher and, cows using sibs plus own record. The intensity after 25 yr, would surpass NMOET. ne~3 of selection (1) for each age group is derived has a long initial lag but also has a high longfrom the P ~ O P O ~ selected O ~ from a normal term response of the foreign population. negenetic responses from the continuous distribution. The selection intensity values for PT1* m* and NMoET are for semen importation strategies are shown in Figfinite population size, and tables in Becker (1) ure 2; p ~ and 1 p ~ results 2 are repeated for are used. comparison. The more that commercial cows Inbreeding is important when Progeny test- are bred from imported semen of moderate ing in a closed PoPulation and in NMoET and price but above average quality, the higher the is not important when imports are involved. response rate is. At 25 yr, the responses for The effects of reduced genetic variance be- CSI30, CSI50, and CSI are 84, 131, and 21 1%, cause of inbreeding are assumed to be negligi- respectively, of that for PTl. ble for the period considered (11) and are The responses from the remaining two ignored. However, the reduction in perfor- strategies are shown in Figure 3; PT1 and PT3 mance because of inbreeding is considered. are presented for comparison. The CEI shows .259
Journal of Dairy Science Vol. 76. No. 4. 1993
1167
GENETIC EVALUATION OF BREEDING STRATEGIES
response 192% of P T I because of a high initial response and a high long-term response. The ELITE strategy suffers from a slower start, but it has a high long-term response rate. In both of these strategies, the young bulls are used untested, and responses exceed PT3 with progeny testing in Zimbabwe. The genetic responses for the nine strategies at yr 25 are presented in Figure 4. Local progeny-testing strategies and import strategies involving low immigration rates rank low. The importation strategies with high gene flow from the exporting country and from high merit parents rank high. The NMOET scheme shows good response, partly because of the high initial genetic lift. These results apply to 25 yr, but the ranking is similar for other periods. These results are for the standard set of Pmm ers. Varying the quality of foreign
4.5
CSI
./' /
0
5
10
I
I
I
15
20
25
Time (yr) 4
,NMOET
3.5
,/' PT2 I
- 3
n
v,
Figure 2. Predicted trends of genetic response in the Zimbabwe commercial cow population for strategies involving continual semen imports. W l = Closed progenytesting scheme; PT2 = progeny testing as PTI but with 30% of the bulls sired by foreign bulls; CSI = continual Semen imports for 30, 50, and 100% of the population for bull to cow pathway.
W
o 2.5
iJ3
C
0
$ 2 a,
L V
.s1.5 e,
c e,
u 1 .5
..., :I .
0
i
I
0
5
10
I
I
I
15
20
25
Time ( y r )
Figure 1. Predicted trends of genetic response in the Zimbabwe commercial cow population for local selection strategies. PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTI but with 30% of the bulls sired by foreign bulls with some imports for the bull to cow pathway; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = scheme using multiple ovulation and embryo transfer nucleus herd (closed).
bulls used for strategies ET2, CSDO, CSI50, and CSI from the top 10% to the top 42% does not have a large effect on genetic response (Table l), possibly because the difference in the genetic superiority among the percentile bull groups is small. The effects of genotype by environment interaction and the initial genetic difference between populations are shown individually in Tables 2 and 3 and jointly in Table 4. When the rg between countries is .5 (Table 2), the ranking of strategies does not differ much from that shown in Figure 4.As rg increases, import strategies become increasingly better than PT1 and NMOET (the strategies based on local selection). When the initial genetic difference between the populations is 0, the strategies involving imports that are competitive are CSI, PT2, ELITE, and CEI (Table 3). The advantage of these strategies over ET1 is reduced substantially. When an NMOET scheme is set Journal of Dairy Science Vol. 76, No. 4, 1993
1168
MPOFU ET AL.
up using animals selected from the Zimbabwe population, NMOET ranks lower than PTl. When the difference between the populations is low (.5 SD), the initial genetic lift for NMOET is low, resulting in a small advantage over pT1. As the initial difference in the mean increases, all strategies, except CSI30, become increasingly better than PTI. When the initial genetic difference is low (.5 SD) and the rg between the two countries is high (rg = l.O), genetic response is higher for all strategies than when the initial genetic difference is high and the rg is low (Table 4), which suggests that the rg between countries should be given more weight than the initial genetic difference between countries when import countries are chosen.
I'.
DISCUSSION
Parameters Used
The different strategies are compared at their maximum. The reproduction rates, selection intensities, and generation intervals used are all high but possible. However, some of the parameters used may be difficult to achieve in practice, resulting in lower response rates. The rate of genetic progress assumed for the exporting population (.I6 SD) is higher than
4.5
I
m c-4
3.5
'
4
" 1
a, L
UJ 3.5
v
.-
-I-'
a,
C
Q,
1
c3
2.5 a, L
n
"
PT1 CS130 CSI PT3 CEI MOET CS150 PT2 ELITE
Strategies
Time (yr) Figure 3. Predicted trends of genetic response in the Zimbabwe commercial cow population for strategies involving imports to produce bulls. PTl = Closed progenytesting scheme; PT3 = progeny testing as PTI but with imports for sire of sons; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos. Journal of Dairy Science Vol. 76. No. 4, 1993
Figure 4. Predicted genetic responses in the Zimbabwe commercial cow population at yr 25 for the nine strategies. PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the bulls sired by foreign bulls with some imports for the bull to cow pathway; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = scheme using multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen impofi for 30% of the population for bull to cow pathway; CSISO = continual semen imports for 50% of the population; CSI = continual semen imports for 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos.
GENETIC EVALUATION OF BREEDING STRATEGIES
rates reported for good progeny-testing schemes, which average about .10 standard deviation (20), although rates (.18 SD) can be higher when progeny testing is combined with importation (5). In comparison, the rate of progress predicted for progeny testing in a closed population for Zimbabwe (F'T1) is .09 standard deviation. The generation intervals used are the same for both populations. The differences in the rate of progress are due to higher selection pressure and to more accurate selection in the exporting population made possible by the larger population size. The generation interval for bull-dams can be reduced by selecting young females, heifers from highly selected progeny-tested bulls, or first lactation cows with high estimated breeding value. The NMOET scheme studied is closed, Christensen (6) demonstrated the superiority of an open herd over a closed herd and its positive effect on genetic gain in a population. Open nucleus schemes have been recommended for developing countries (10). In the present study, NMOET improves on PT1 by 66% at 25 yr. In a study by Woolliams and Smith (22), adult MOET schemes improved on conventional progeny-testing schemes by 20%. In their case (22), the initial stock for the nucleus herd was selected from the local population rather than from a foreign population, the cow population of which is, on average, 1.25 standard deviation better than the local population. The genetic difference between elite cows and the commercial cow population within the same population is approximately .8
1169
standard deviation, and the difference between sires of replacements and sires of sons is .081 standard deviation (20). The initial genetic lift for MOET schemes in this case would not be high. When the initial genetic difference between the populations is .5 standard deviation, the NMOET scheme is only 33% better than PT1. Deficiences in Evaluation Model
Selection of parents gives rise to a temporary disequilibrium and to less genetic variation for selection among offspring (4). For this study, the two populations are considered to have been under continuous selection for milk yield and to have reached selection equilibrium. Therefore, the genetic parameters used already include the reduction from the Bulmer effect. With NMOET schemes, bull evaluations are on information on collateral relatives, resulting in a high correlation of estimated breeding values of family members, which, in turn, would result in smaller effective selection differentials in small populations (8). For simplicity, the P matrix is assumed to be constant over the evaluation period for each strategy; i.e., a priori assumptions are made about genetic means for each age group and generation intervals. With truncation selection across distributions, these terms are not easy to quantify a priori because their values depend on genetic differences among age groups and on the previous selection rounds. Keeping P fixed introduces error for the expected genera-
TABLE l . The effect on response of the quality of foreign bulls used in Zimbabwe for strategies based on continual semen importation for the bulls to breed cows pathway.
TOP
Genetic superiority of sire group
(%)
SD
10
.7? .62 .51 .41
20 30 42
Genetic response at yr 25 (SD)
PT2'
CS130
CSISO
CSI
3.32 3.32 3.31 3.31
1.79 1.79 1.78 1.78
2.80 2.80 2.79 2.78
4.50 4.49 4.48 4.47
'PT2= Progeny testing with some impom for the bull to cow pathway; CSI30 = continual semen importation for 30% of the population, CSISO = continual Semen importation for 50% of the population, CSI = continual semen importation for 100% of the population. Journal of Dairy Science Vol. 76, No. 4, 1993
1170
MPOFU ET AL.
TABLE 2. The effect of genotype by environment interaction, measured by the genetic correlation (rg). on response for nine strategies when the initial differences between the populations is 1.25 standard deviations. Genetic response at yr 25 (SD) rs Strategy' PTl NMOET CSI30 CSI50 CSI PT2 PT3 ELITE strategy CEI strategy
.5
.7
.9
1.0
2.12 3.51 1S O 2.35 3.73 3.01 2.79 3.03 3.44
2.12 3.51 1.78 2.78 4.47 3.31 3.02 3.41 4.07
2.12 3.51 2.05 3.22 5.21 3.61 3.26 3.80 4.70
2.12 3.51 2.19 3.43 5.58 3.76 3.37 3.99 5.01
'PTl = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as P T I but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation for 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos.
tion interval, which overestimates genetic response (11). When the purpose of a study is not to obtain accurate estimates of genetic means but to determine relative means for the strategies, a changing P matrix is not necessary (3). Implications of the Results
In Zimbabwe, before the introduction of a progeny-testing scheme in 1987, farmers relied
on semen importation for genetic improvement. Approximately 30% of heifers were sired by foreign bulls (13). This strategy is similar to CSI30. Only limited progress was made, and the slow progress was explained as being due to lack of well-defined national breeding goals (23). From the results of this study, the slow progress was partly due to the strategy chosen. Smith (19) suggested a breeding strategy with an initial period of semen importation and a later switch to a local selec-
TABLE 3. The effect of the magnitude of the initial difference in genetic means between the two populations on response when the genetic correlation between the populations is .7. Genetic response at yr 25 (SD) Initial genetic difference Strategy' PTl NMOET CSI30 CSI50 CSI PT2 PT3 ELITE CEI
0
2.12 1.79 .95 1.52 2.59 2.48 2.11 2.23 2.19
.50
1.00
1.25
1S O
2.12 2.82 1.46 2.29 3.73 2.99 2.71 2.99 3.34
2.12 3.33 1.69 2.65 4.28 3.22 2.92 3.29 3.88
2.12 3.51 1.78 2.78 4.47 3.31 3.02 3.41 4.07
2.12 3.74 1.88 2.94 4.71 3.41 3.12 3.55 4.30
'PTl = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation of 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos. Journal of Dairy Science Vol. 76, No. 4. 1993
1171
GENETIC EVALUATION OF BREEDING STRATEGIES
TABLE 4. The effect of the joint action of the initial genetic differences (GD) and genetic correlation (rg) between countries on response for strategies based on importation. Genetic response at yr 25 Strategy1 CSI30 CSI50 CSI PT2 PT3 ELITE
CEI
GD = 1.5 SD, rg = .5
GD = .5 SD, rg = 1.0
GD = 1.25 SD, rg = .7
SD 1.61 2.50 3.97 3.11 2.88 3.17 3.68
SD
SD 1.78 2.78 4.47 3.31 3.02 3.41 4.07
1.87 2.94 4.84 3.43 3.07 3.57 4.28
_ _ _ _ _ _ _ _ ~
~
'PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTI but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as PTI but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CS130 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation of 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos
tion program. This study shows that CSDO is better than PTl for the first 20 yr of the evaluation period (Figure 1). Therefore, CSBO was better in the initial period than using a progeny-testing scheme initially. The results have shown that an even better strategy would have been to import embryos to produce bulls for the commercial cow population. Currently in Zimbabwe some of the bulls on test are sons of foreign bulls, and some of the semen imported is for the bull to cow pathway, a combination of strategies PT2 and PT3. Because of foreign currency restrictions imposed by the government, only small amounts of semen are imported. Therefore, the current strategy is more inclined toward PT3 than PT2. The PT2 gives a 9% higher genetic mean than PT3 at 25 yr. The choice of strategy will depend on the value of this difference and on the price of imported semen from elite bulls compared with that of other selected bulls. An NMOET scheme has never been used in Zimbabwe. The NMOET is the best local strategy and ranks closely to the best strategies based on imports. The success of NMOET will depend on reproductive rates achieved in Zimbabwe. The rate of progress for strategies based on imports depends on the immigration rate and the initial difference in the genetic mean between the exporting and the importing popula-
tion. Zimbabwe should import from countries in which genetic means are higher. If, for financial or other reasons, imports can only be from populations with a low mean (though not lower than Zimbabwe's), importation strategies involving low usage of foreign genetic material (CSI30) should be avoided. Genotype by environment interaction between countries is important in determining the rate of progress in the importing country and should be given more consideration than the initial genetic difference between the populations. Strategies based on imports involve the use of scarce foreign currency. In addition, the differences in predicted responses between some of the strategies are small. An economic evaluation of the strategies is needed to determine whether the genetic differences among strategies lead to greater economic benefits. REFERENCES 1 Becker, W. A. 1985. Manual of Quantitative Genetics.
4th ed. Academic Press, Pullman, WA. 2 Bichard, M. 1971. Dissemination of genetic improvement through a livestock industry. Anim. Prod. 13: 401. 3 Brascamp, E. W. 1978. Methods on economic optimization of animal breeding plans. Rep. B-134. Res. Inst. Anim. Husbandry, Zeist, Neth. 4Bulmer. M. G. 1971. The effect of selection on genetic variability. Am. Nat. 105:210. 5 Bumside, E. B., G. B. Jansen, G. Civati, and E.
Journal of Dairy Science Vol. 76, No. 4, 1993
1172
MPOFU ET AL.
Dadati. 1992. Observed and theoretical genetic trends in a large dairy population under intensive selection. J. Dairy Sci. 75:2242. 6 Christensen. L. G. 1987. MOET programmes in cattle breeding. Proc. 38th Annu. Mtg. Eur. Assoc. Anim. Prod., Lisbon, Portugal. 7Falconer. D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. Longmans. Essex, Engl. 8 Hill, W. G. 1971. Design and efficiency of selection experiments for estimating genetic parameters. Biometrics 27:293. 9Hil1, W. G. 1974. Prediction and evaluation of response to selection with overlapping generations. Anim. Prod. 18:117. 10 Hodges, J. 1991. Genetic improvement of livestock in developing countries using the open nucleus breeding system. Proc. FA0 Conf. Open Nucleus Breeding Systems, Bialobrzegi, Poland. Food Agric. Org., Rome, Italy. 11 Meuwissen, T.H.E. 1989. A deterministic model for the optimisation of dairy cattle breeding schemes with special emphasis on selection parameters. Anim. Rod. 49:193. 12Miglior. F., B. Szkotnicki, and E.B. Burnside. 1992. Analysis of levels of inbreeding and inbreeding depression in Jersey cattle. J. Dairy Sci. 75:1112. 13 Mpofu, N. 1986. Genetic parameters for milk production in Zimbabwean black and white cows. M.Sc. Thesis, Univ. Edinburgh, Edinburgh, Scotland.
Journal of Dairy Science Vol. 76, No. 4, 1993
14 Mpofu, N. 1992. Genetic and economic evaluation of dairy cattle breeding strategies for Zimbabwe. Ph.D. Diss., Univ. Guelph, Guelph, ON, Can. 15Mpofu, N., C. Smith, W. Van Vuuren, and E. B. Burnside. 1993. Breeding strategies for genetic improvement of dairy cattle in Zimbabwe. 2. Economic evaluation. J. Dairy Sci. 76:1173. 16Pirchner, F. 1983. Population Genetics in Animal Breeding. 2nd ed. Plenum Press, New York, NY. 17 Rendel, J. M.,and A. Robertson. 1950. Estimation of genetic gain in milk yield by selection in a closed herd of dairy cattle. J. Genet. 50:l. 18Ruane. J., and R. Thompson. 1991. Comparison of simulated and theoretical results in adult MOET schemes for dairy cattle. Livest. Prod. Sci. 28:l. 19Smith. C. 1988. Breeding strategies for developing countries. World Anim. Rev. 65:2. 20Van Tassell, C. P., and L. D. Van Vleck. 1991. Estimates of genetic selection differentials and generation intervals for the four paths of selection. J. Dairy Sci. 74:1078. 21 Van Vleck, L. D. 1986. Evaluation of dairy cattle breeding programs: specialised milk production. Proc. 3rd World Congr. Genet. Appl. Livest. Prod., Lincoln, NE IX:144. 22 Woolliams, J. A,, and C. Smith. 1988. The value of indicator traits in the genetic improvement of dairy cattle. Anim. Prod. 46:333.
Abbreviation key: CEI strategy = untested young bull team produced from continual embryo importation; CSI = continual semen importation for 100% of the population; CSI30 = continual semen importation for 30% of the population; CSISO = continual semen importation for 50% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; MOET = multiple ovulation and embryo transfer; NMOET = scheme using MOET nucleus herd (closed); PT1 = closed progenytesting scheme; FT2 = progeny testing as PT1 but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as FT1 but with imports for sire of sons; rg = genetic correlation of performance between countries.
Strategies of selection within local populations and of importation of semen and embryos from countries with higher genetic merit are evaluated for genetic improvement of milk yield in Zimbabwe. Local programs are progeny testing in a closed population, progeny testing combined with semen importation to sire 30% of cows, progeny testing with foreign sires as sires of bulls, and a closed nucleus selection scheme using embryo transfer started from elite imported stock. Strategies based on continuous importation are imported semen for 30, 50, or 100% of cows and semen from elite foreign bulls used on local elite cows and bulls from elite imported embryos. Gene flow methods were used to estimate the genetic means for the strategies. In predicted genetic response at 25 yr, continuous importation of all semen from the superior exporting country ranks first; second is the importation of elite embryos to provide bulls for AI; and third is the closed nucleus scheme using embryo transfer. Schemes involving importation of semen to breed bulls with local progeny testing are intermediate, and the lowest responses come from semen imported to breed cows and from the closed local progeny-testing scheme. Ranking of the strategies is influenced by the genetic correlation of performance and by the initial genetic difference between Zimbabwe and the exporting country. (Key words: local selection, imports, breeding strategies)
INTRODUCTION
Received July 24, 1992. Accepted November 9, 1992. 'Current address: Department of Animal Science. University of Zimbabwe, Harare, Zimbabwe. 1993 J Dairy Sci 76:1163-1172
The two main methods of genetic improvement are selection and migration. Selection techniques, such as progeny testing, used in most developed countries should be applicable to Zimbabwe because genetic parameters for milk yield are similar (13) and herd recording is increasing. The rates of progress made through progeny testing, although low, are cumulative, thus resulting in improved milk yield over time. Given the costs of such selection programs, Smith (19) suggested that developing countries adopt breeding strategies based on importation or on nucleus breeding schemes using multiple ovulation and embryo transfer (MOET). Imports are recommended when 1) exotic stocks are genetically better than domestic stock, 2) when selection objectives are similar, and 3) when genotype by environment interaction is not important. In Zimbabwe, temperate breeds are used for commercial dairy production. The breeding objective in Zimbabwe is to increase genetic merit for milk yield. The 1990 national average yield was 5830 kg per cow per 300-d lactation (23).
1163
1164
MPOFU ET AL.
The Zimbabwe Holstein population was estimated to be genetically 1.25 SD [phenotypic (1 SD = 540 kg)] below the US population, and the genotype by environment interaction between the two countries was estimated by a genetic correlation of .70 f .18 (14). Therefore, strategies involving importation of genetic material from the US or other temperate countries should be evaluated. The objective in this paper is to evaluate genetically breeding strategies for Zimbabwe. Four strategies based on domestic selection programs and five based on importation of foreign genetic material are considered. The effects of genotype by environment interaction and the initial difference in genetic means between Zimbabwe and the exporting population on the ranking of the strategies are also studied. A complete evaluation of strategies should include an economic evaluation, which is the subject of a companion paper (15). MATERIALS AND METHODS
The parameters assumed for the Zimbabwe dairy population are a population of 100,OOO cows with 40% milk recorded, and, of these, 80% are bred by AI (Zimbabwe Milk Record Office, Harare, Zimbabwe, 1984 to 1990, annual and internal reports). Heifers calve for the first time at 24 mo of age; calving interval is 12 mo. Cows may survive up to five lactations, and the distribution by lactation number is 30, 25, 20, 15, and 10% for lactations 1 to 5, respectively. For simplicity, imports are assumed to be from one country, the US. The exporting country is assumed to have a very large dairy population, to have a long-standing, closed progeny-testing scheme with an annual rate of progress of .16 standard deviation, and to have selected cattle that are genetically superior to Zimbabwe cattle. Assumptlons for the Proposed Breeding Strategies
Progeny Testing. Three progeny-testing strategies are studied. The first strategy (PT1) is a closed progeny-testing scheme. Thirty-five young bulls are tested per year on 40% of the recorded cows bred by AI. Each test bull has 40 daughters with completed lactations. The top 3% of test bulls are selected as sires of Journal of Dairy Science Vol. 76, No. 4. 1993
sons and the top 9% as sires of replacements, each bull producing 40,000 doses of semen. The generation intervals are 2 yr for the test bulls, 6 yr for sires of replacements, and 7 yr for sires of sons. Only cows in their third and fourth lactations (ages 4 and 5 yr) are selected as bull-dams, and these two parities contribute equally to the next generation. Eight bull-dams are selected for every test bull required. No selection is practiced for dams of cows. The generation interval for the cows is 4 yr. The second strategy (PT2) differs from PT1 in that 30% of the cow population is sired by foreign bulls. Thus, a smaller bull stud than for PT1 is required. Only 20% of the cows are bred to test bulls, and each test bull has 20 daughters with complete lactations. The third strategy (PT3) is like PT1, but the best available foreign bulls are used as sires of sons. These bulls are all 6 yr old. MOET Nucleus Scheme. A MOET nucleus scheme (NMOET) is set up to produce and to test bulls for use in the commercial cow population. The nucleus herd is founded by screening and selecting stock with the highest estimated breeding values from the exporting population. The initial genetic mean of the nucleus is therefore higher than that of the Zimbabwean base population. Thereafter, selection is within the nucleus. The size of the herd is 50 donor cows. Donor cows are superovulated, giving 8 embryos per flush. With a pregnancy rate of 65% (obtained for fresh embryos in Zimbabwe), a calf survival rate of 80%, and a sex ratio of 50%, the average number of surviving male (and female) offspring per flush is 2. The cows are flushed four times and are mated to a different bull at each flush. Approximately 4 bulls are required to service the base (commercial breeding) population. Because the MOET bulls are selected on sibling information, selection accuracy is lower, so the rate of use per bull is reduced, and twice as many bulls as required are selected for the base population. A restriction of 1 bull selected per full sibship and 1 per dam results in a selection pressure of 4 out of 50 for nucleus sires and 8 out of 50 for sires for the base population. The selection pressure for nucleus females is 50 out of 200. The generation interval is 51 mo. Selection of bulls is on their full and half-sib sisters’ first lactation records.
1165
GENETIC EVALUATION OF BREEDING STRATEGIES
Continual Semen Importation. With continual semen importation, there would be no bull testing program in Zimbabwe. Semen from above average proven bulls (as opposed to elite bulls) is imported to sire replacements. The proportion of cows sired by foreign bulls is varied to include 30 (CSDO), 50 (CSISO), and 100% (CSI). The imports are made every year with each new batch of foreign proven sire semen higher in genetic merit than the previous batch because of genetic trend. When less than 100% of the cows are bred to foreign bulls, the rest of the cows are bred to untested local bulls. Local bulls are used from 2 to 6 yr of age, and the age groups contribute equally to the next generation. Even when semen is imported mainly to sire cows, some of the local bulls are sired by foreign bulls, contributing 10% of the untested local bulls for CSI30 and 20% for the CSISO strategy. Untested Young Bull Teams from Local Elite Cows and Elite Foreign Bulls. The ELITE strategy is based on continued importation of semen from elite bulls to sire sons. The young bulls are used untested. To reduce the risk from untested bulls, the rate of usage per bull is reduced from the 40,000 AI for progeny-tested bulls to 4000 AI per bull, so 40 young bulls are required. Bull-dams are elite cows selected from the local population; MOET is used, making it possible to produce at least 1 bull per bull-dam. Embryo collection is conducted in individual herds; therefore, embryo yields are expected to be lower than for NMOET. Pregnancy rates should be similar to those for NMOET because in both cases fresh embryos are used. Semen collection starts at 15 mo, and bulls are used for 1 yr. The female calves from elite matings are kept in the herds in which they are born and can be selected as bull-dams. Continual Embtyo Importation. In the CEI strategy, highly selected bull-dams in the exporting country are mated to elite bulls to produce embryos for export. Young bulls from these embryos are used untested for regular service at 15 mo of age, and the heifers are sold to commercial herds. As for the ELITE strategy, 40 young bulls are produced for the CEI strategy. Because the embryos will be frozen, the success rate (32%) is lower.
Method of Evaluation
The effectiveness of the different breeding strategies is assessed by comparing predicted genetic progress possible by adoption of each strategy using deterministic simulation methods. Rendel and Robertson (17) predicted genetic change based on four selection pathways: bulls to breed bulls, bulls to breed cows, cows to breed bulls, and cows to breed cows. The estimate of annual genetic change (AGy) is
AG,, = CAGEL
[I1
where AG is the genetic superiority for each of the four pathways, and L is the generation interval for each of the four pathways. The equation gives asymptotic response rates for overlapping generations. Hill (9) presented a gene flow procedure using matrix formulation based on the Rendel and Robertson (17) formula. These procedures are used to evaluate the breeding strategies. The evaluation criterion is the weighted genetic mean for cows in their first to fifth lactations, weighting by the proportion of the age group. The strategies are evaluated over a 25-yr period. The gene flow model used is
M(t) = P[M(t - 1)
+ AG(t)l
[2]
where M(t) is a vector of average genetic merit of animals at time t, P is a gene transmission matrix, and M(t - 1) is a vector of average genetic merit of animals at time (t - l), and AG(t) is a vector of the genetic superiority of selected animals. Continuous selection in both sexes is assumed. The P matrix specifies the contribution of different age classes to the next generation. The P matrix is assumed to be fixed for each strategy; i.e., the contributions of the different age classes to the next generation are assumed to be the same across years. The elements of the P matrix are defined to reach assigned generation intervals for each strategy. The initial average genetic merit [M(t = O)] is set to 0 for the Zimbabwe cow population. The genetic difference between the Zimbabwe cow population and that of the importing country is assumed to be 1.25 standard deviations (675 kg of milk). This value is used for the initial genetic difference between the NMOET and the Zimbabwean base population. Imports Journal of Dairy Science Vol. 76, No. 4, 1993
1166
MFQFU ET AL.
are either semen from tested bulls or embryos sired by elite bulls. Cows are assumed to lag behind selected bulls by 3AG (2, 15), where AG is the assumed annual genetic response. For the exporting country, the lag is .48 standard deviations (.16 x 3). The mean for the selected foreign bulls is therefore equal to 1.73 standard deviations (1.25 + .48). Elite bulls are assumed to be the top 10%. Semen from these bulls is imported for strategies PT3 and ELITE, and embryos for CEI are sired by elite bulls. Semen for strategies PT2, CSDO, CSI50, and CSI is from bulls that are .21 standard deviation above the mean or the top 42% bulls (14). The importation of semen from bulls in the top 10, 20, and 30% is investigated for pT2, CSDO, CSI50, and CSI. The within-group selection differential of the selected animals at time t, AG(i,t), is given by
where rg in milk yield is between the importing and exporting countries. The effect of rg on the ranking of breeding strategies is investigated. The range of values used is .5, .7, .9, and 1.0. The rg between performance in the nucleus and the base population is assumed to be 1.0 for NMOET.
AG = r x I x ug,
RESULTS
[31
The rate of inbreeding is calculated using Wright's formula (7). The inbreeding rates are scaled up by a factor of 2 for NMOET following Ruane and Thompson (18). The reduction in the mean per year because of inbreeding depression is calculated assuming .27% depression in milk yield per percentage of increase in inbreeding (12). The genetic superiority for imports to Zimbabwe is calculated as AG = I x r x u, x rg
[41
The genetic responses are calculated over where I is the intensity of selection, r is the years up to yr 25, using a standard set of and ug is the genetic accuracy Of parameters (initial difference, 1.25 SD, rg = standard deviation for milk yield. Heritability .7), and refer to the Zimbabwe commercial for Yield was assumed to be genetic cow population. The responses are then corncorrelation (r,) to be 1.00 between successive pared with pT1, the closed Zimbabwe lactations, and repeatability of yield over SUC- progeny-testing scheme. varicessive lactations to be e40. The results for the local selection strategies ance is standardized to 1.00. The accuracy of are in Figure 1. The response is highest with evaluation (r) for both bulls and COWS is calcu- NMoET: 66% higher than pT1 at yr 25, lated using selection index theory for bulls largely because of the initial genetic lift using daughter information and for cows using recouped by NMOET for the long-term genetic information from the she, the dam, and the response rates are similar. with p ~ 2in, which cow*s Own lactation For the "loET 30% of cows are bred by foreign bulls, an scheme, the accuracy of evaluation for bulls is early (but smaller) response is also obtained, calculated using sibling information and for but the long-term response rate is higher and, cows using sibs plus own record. The intensity after 25 yr, would surpass NMOET. ne~3 of selection (1) for each age group is derived has a long initial lag but also has a high longfrom the P ~ O P O ~ selected O ~ from a normal term response of the foreign population. negenetic responses from the continuous distribution. The selection intensity values for PT1* m* and NMoET are for semen importation strategies are shown in Figfinite population size, and tables in Becker (1) ure 2; p ~ and 1 p ~ results 2 are repeated for are used. comparison. The more that commercial cows Inbreeding is important when Progeny test- are bred from imported semen of moderate ing in a closed PoPulation and in NMoET and price but above average quality, the higher the is not important when imports are involved. response rate is. At 25 yr, the responses for The effects of reduced genetic variance be- CSI30, CSI50, and CSI are 84, 131, and 21 1%, cause of inbreeding are assumed to be negligi- respectively, of that for PTl. ble for the period considered (11) and are The responses from the remaining two ignored. However, the reduction in perfor- strategies are shown in Figure 3; PT1 and PT3 mance because of inbreeding is considered. are presented for comparison. The CEI shows .259
Journal of Dairy Science Vol. 76. No. 4. 1993
1167
GENETIC EVALUATION OF BREEDING STRATEGIES
response 192% of P T I because of a high initial response and a high long-term response. The ELITE strategy suffers from a slower start, but it has a high long-term response rate. In both of these strategies, the young bulls are used untested, and responses exceed PT3 with progeny testing in Zimbabwe. The genetic responses for the nine strategies at yr 25 are presented in Figure 4. Local progeny-testing strategies and import strategies involving low immigration rates rank low. The importation strategies with high gene flow from the exporting country and from high merit parents rank high. The NMOET scheme shows good response, partly because of the high initial genetic lift. These results apply to 25 yr, but the ranking is similar for other periods. These results are for the standard set of Pmm ers. Varying the quality of foreign
4.5
CSI
./' /
0
5
10
I
I
I
15
20
25
Time (yr) 4
,NMOET
3.5
,/' PT2 I
- 3
n
v,
Figure 2. Predicted trends of genetic response in the Zimbabwe commercial cow population for strategies involving continual semen imports. W l = Closed progenytesting scheme; PT2 = progeny testing as PTI but with 30% of the bulls sired by foreign bulls; CSI = continual Semen imports for 30, 50, and 100% of the population for bull to cow pathway.
W
o 2.5
iJ3
C
0
$ 2 a,
L V
.s1.5 e,
c e,
u 1 .5
..., :I .
0
i
I
0
5
10
I
I
I
15
20
25
Time ( y r )
Figure 1. Predicted trends of genetic response in the Zimbabwe commercial cow population for local selection strategies. PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTI but with 30% of the bulls sired by foreign bulls with some imports for the bull to cow pathway; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = scheme using multiple ovulation and embryo transfer nucleus herd (closed).
bulls used for strategies ET2, CSDO, CSI50, and CSI from the top 10% to the top 42% does not have a large effect on genetic response (Table l), possibly because the difference in the genetic superiority among the percentile bull groups is small. The effects of genotype by environment interaction and the initial genetic difference between populations are shown individually in Tables 2 and 3 and jointly in Table 4. When the rg between countries is .5 (Table 2), the ranking of strategies does not differ much from that shown in Figure 4.As rg increases, import strategies become increasingly better than PT1 and NMOET (the strategies based on local selection). When the initial genetic difference between the populations is 0, the strategies involving imports that are competitive are CSI, PT2, ELITE, and CEI (Table 3). The advantage of these strategies over ET1 is reduced substantially. When an NMOET scheme is set Journal of Dairy Science Vol. 76, No. 4, 1993
1168
MPOFU ET AL.
up using animals selected from the Zimbabwe population, NMOET ranks lower than PTl. When the difference between the populations is low (.5 SD), the initial genetic lift for NMOET is low, resulting in a small advantage over pT1. As the initial difference in the mean increases, all strategies, except CSI30, become increasingly better than PTI. When the initial genetic difference is low (.5 SD) and the rg between the two countries is high (rg = l.O), genetic response is higher for all strategies than when the initial genetic difference is high and the rg is low (Table 4), which suggests that the rg between countries should be given more weight than the initial genetic difference between countries when import countries are chosen.
I'.
DISCUSSION
Parameters Used
The different strategies are compared at their maximum. The reproduction rates, selection intensities, and generation intervals used are all high but possible. However, some of the parameters used may be difficult to achieve in practice, resulting in lower response rates. The rate of genetic progress assumed for the exporting population (.I6 SD) is higher than
4.5
I
m c-4
3.5
'
4
" 1
a, L
UJ 3.5
v
.-
-I-'
a,
C
Q,
1
c3
2.5 a, L
n
"
PT1 CS130 CSI PT3 CEI MOET CS150 PT2 ELITE
Strategies
Time (yr) Figure 3. Predicted trends of genetic response in the Zimbabwe commercial cow population for strategies involving imports to produce bulls. PTl = Closed progenytesting scheme; PT3 = progeny testing as PTI but with imports for sire of sons; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos. Journal of Dairy Science Vol. 76. No. 4, 1993
Figure 4. Predicted genetic responses in the Zimbabwe commercial cow population at yr 25 for the nine strategies. PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the bulls sired by foreign bulls with some imports for the bull to cow pathway; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = scheme using multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen impofi for 30% of the population for bull to cow pathway; CSISO = continual semen imports for 50% of the population; CSI = continual semen imports for 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos.
GENETIC EVALUATION OF BREEDING STRATEGIES
rates reported for good progeny-testing schemes, which average about .10 standard deviation (20), although rates (.18 SD) can be higher when progeny testing is combined with importation (5). In comparison, the rate of progress predicted for progeny testing in a closed population for Zimbabwe (F'T1) is .09 standard deviation. The generation intervals used are the same for both populations. The differences in the rate of progress are due to higher selection pressure and to more accurate selection in the exporting population made possible by the larger population size. The generation interval for bull-dams can be reduced by selecting young females, heifers from highly selected progeny-tested bulls, or first lactation cows with high estimated breeding value. The NMOET scheme studied is closed, Christensen (6) demonstrated the superiority of an open herd over a closed herd and its positive effect on genetic gain in a population. Open nucleus schemes have been recommended for developing countries (10). In the present study, NMOET improves on PT1 by 66% at 25 yr. In a study by Woolliams and Smith (22), adult MOET schemes improved on conventional progeny-testing schemes by 20%. In their case (22), the initial stock for the nucleus herd was selected from the local population rather than from a foreign population, the cow population of which is, on average, 1.25 standard deviation better than the local population. The genetic difference between elite cows and the commercial cow population within the same population is approximately .8
1169
standard deviation, and the difference between sires of replacements and sires of sons is .081 standard deviation (20). The initial genetic lift for MOET schemes in this case would not be high. When the initial genetic difference between the populations is .5 standard deviation, the NMOET scheme is only 33% better than PT1. Deficiences in Evaluation Model
Selection of parents gives rise to a temporary disequilibrium and to less genetic variation for selection among offspring (4). For this study, the two populations are considered to have been under continuous selection for milk yield and to have reached selection equilibrium. Therefore, the genetic parameters used already include the reduction from the Bulmer effect. With NMOET schemes, bull evaluations are on information on collateral relatives, resulting in a high correlation of estimated breeding values of family members, which, in turn, would result in smaller effective selection differentials in small populations (8). For simplicity, the P matrix is assumed to be constant over the evaluation period for each strategy; i.e., a priori assumptions are made about genetic means for each age group and generation intervals. With truncation selection across distributions, these terms are not easy to quantify a priori because their values depend on genetic differences among age groups and on the previous selection rounds. Keeping P fixed introduces error for the expected genera-
TABLE l . The effect on response of the quality of foreign bulls used in Zimbabwe for strategies based on continual semen importation for the bulls to breed cows pathway.
TOP
Genetic superiority of sire group
(%)
SD
10
.7? .62 .51 .41
20 30 42
Genetic response at yr 25 (SD)
PT2'
CS130
CSISO
CSI
3.32 3.32 3.31 3.31
1.79 1.79 1.78 1.78
2.80 2.80 2.79 2.78
4.50 4.49 4.48 4.47
'PT2= Progeny testing with some impom for the bull to cow pathway; CSI30 = continual semen importation for 30% of the population, CSISO = continual Semen importation for 50% of the population, CSI = continual semen importation for 100% of the population. Journal of Dairy Science Vol. 76, No. 4, 1993
1170
MPOFU ET AL.
TABLE 2. The effect of genotype by environment interaction, measured by the genetic correlation (rg). on response for nine strategies when the initial differences between the populations is 1.25 standard deviations. Genetic response at yr 25 (SD) rs Strategy' PTl NMOET CSI30 CSI50 CSI PT2 PT3 ELITE strategy CEI strategy
.5
.7
.9
1.0
2.12 3.51 1S O 2.35 3.73 3.01 2.79 3.03 3.44
2.12 3.51 1.78 2.78 4.47 3.31 3.02 3.41 4.07
2.12 3.51 2.05 3.22 5.21 3.61 3.26 3.80 4.70
2.12 3.51 2.19 3.43 5.58 3.76 3.37 3.99 5.01
'PTl = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as P T I but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation for 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos.
tion interval, which overestimates genetic response (11). When the purpose of a study is not to obtain accurate estimates of genetic means but to determine relative means for the strategies, a changing P matrix is not necessary (3). Implications of the Results
In Zimbabwe, before the introduction of a progeny-testing scheme in 1987, farmers relied
on semen importation for genetic improvement. Approximately 30% of heifers were sired by foreign bulls (13). This strategy is similar to CSI30. Only limited progress was made, and the slow progress was explained as being due to lack of well-defined national breeding goals (23). From the results of this study, the slow progress was partly due to the strategy chosen. Smith (19) suggested a breeding strategy with an initial period of semen importation and a later switch to a local selec-
TABLE 3. The effect of the magnitude of the initial difference in genetic means between the two populations on response when the genetic correlation between the populations is .7. Genetic response at yr 25 (SD) Initial genetic difference Strategy' PTl NMOET CSI30 CSI50 CSI PT2 PT3 ELITE CEI
0
2.12 1.79 .95 1.52 2.59 2.48 2.11 2.23 2.19
.50
1.00
1.25
1S O
2.12 2.82 1.46 2.29 3.73 2.99 2.71 2.99 3.34
2.12 3.33 1.69 2.65 4.28 3.22 2.92 3.29 3.88
2.12 3.51 1.78 2.78 4.47 3.31 3.02 3.41 4.07
2.12 3.74 1.88 2.94 4.71 3.41 3.12 3.55 4.30
'PTl = Closed progeny-testing scheme; PT2 = progeny testing as PTl but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as PTl but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CSI30 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation of 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos. Journal of Dairy Science Vol. 76, No. 4. 1993
1171
GENETIC EVALUATION OF BREEDING STRATEGIES
TABLE 4. The effect of the joint action of the initial genetic differences (GD) and genetic correlation (rg) between countries on response for strategies based on importation. Genetic response at yr 25 Strategy1 CSI30 CSI50 CSI PT2 PT3 ELITE
CEI
GD = 1.5 SD, rg = .5
GD = .5 SD, rg = 1.0
GD = 1.25 SD, rg = .7
SD 1.61 2.50 3.97 3.11 2.88 3.17 3.68
SD
SD 1.78 2.78 4.47 3.31 3.02 3.41 4.07
1.87 2.94 4.84 3.43 3.07 3.57 4.28
_ _ _ _ _ _ _ _ ~
~
'PTI = Closed progeny-testing scheme; PT2 = progeny testing as PTI but with 30% of the cows sired by foreign bulls; PT3 = progeny testing as PTI but with imports for sire of sons; NMOET = multiple ovulation and embryo transfer nucleus herd (closed); CS130 = continual semen importation for 30% of the population; CSI50 = continual semen importation for 50% of the population; CSI = continual semen importation of 100% of the population; ELITE strategy = untested young bull team produced from foreign elite bulls and local elite cows; CEI strategy = untested young bull team produced from imported embryos
tion program. This study shows that CSDO is better than PTl for the first 20 yr of the evaluation period (Figure 1). Therefore, CSBO was better in the initial period than using a progeny-testing scheme initially. The results have shown that an even better strategy would have been to import embryos to produce bulls for the commercial cow population. Currently in Zimbabwe some of the bulls on test are sons of foreign bulls, and some of the semen imported is for the bull to cow pathway, a combination of strategies PT2 and PT3. Because of foreign currency restrictions imposed by the government, only small amounts of semen are imported. Therefore, the current strategy is more inclined toward PT3 than PT2. The PT2 gives a 9% higher genetic mean than PT3 at 25 yr. The choice of strategy will depend on the value of this difference and on the price of imported semen from elite bulls compared with that of other selected bulls. An NMOET scheme has never been used in Zimbabwe. The NMOET is the best local strategy and ranks closely to the best strategies based on imports. The success of NMOET will depend on reproductive rates achieved in Zimbabwe. The rate of progress for strategies based on imports depends on the immigration rate and the initial difference in the genetic mean between the exporting and the importing popula-
tion. Zimbabwe should import from countries in which genetic means are higher. If, for financial or other reasons, imports can only be from populations with a low mean (though not lower than Zimbabwe's), importation strategies involving low usage of foreign genetic material (CSI30) should be avoided. Genotype by environment interaction between countries is important in determining the rate of progress in the importing country and should be given more consideration than the initial genetic difference between the populations. Strategies based on imports involve the use of scarce foreign currency. In addition, the differences in predicted responses between some of the strategies are small. An economic evaluation of the strategies is needed to determine whether the genetic differences among strategies lead to greater economic benefits. REFERENCES 1 Becker, W. A. 1985. Manual of Quantitative Genetics.
4th ed. Academic Press, Pullman, WA. 2 Bichard, M. 1971. Dissemination of genetic improvement through a livestock industry. Anim. Prod. 13: 401. 3 Brascamp, E. W. 1978. Methods on economic optimization of animal breeding plans. Rep. B-134. Res. Inst. Anim. Husbandry, Zeist, Neth. 4Bulmer. M. G. 1971. The effect of selection on genetic variability. Am. Nat. 105:210. 5 Bumside, E. B., G. B. Jansen, G. Civati, and E.
Journal of Dairy Science Vol. 76, No. 4, 1993
1172
MPOFU ET AL.
Dadati. 1992. Observed and theoretical genetic trends in a large dairy population under intensive selection. J. Dairy Sci. 75:2242. 6 Christensen. L. G. 1987. MOET programmes in cattle breeding. Proc. 38th Annu. Mtg. Eur. Assoc. Anim. Prod., Lisbon, Portugal. 7Falconer. D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. Longmans. Essex, Engl. 8 Hill, W. G. 1971. Design and efficiency of selection experiments for estimating genetic parameters. Biometrics 27:293. 9Hil1, W. G. 1974. Prediction and evaluation of response to selection with overlapping generations. Anim. Prod. 18:117. 10 Hodges, J. 1991. Genetic improvement of livestock in developing countries using the open nucleus breeding system. Proc. FA0 Conf. Open Nucleus Breeding Systems, Bialobrzegi, Poland. Food Agric. Org., Rome, Italy. 11 Meuwissen, T.H.E. 1989. A deterministic model for the optimisation of dairy cattle breeding schemes with special emphasis on selection parameters. Anim. Rod. 49:193. 12Miglior. F., B. Szkotnicki, and E.B. Burnside. 1992. Analysis of levels of inbreeding and inbreeding depression in Jersey cattle. J. Dairy Sci. 75:1112. 13 Mpofu, N. 1986. Genetic parameters for milk production in Zimbabwean black and white cows. M.Sc. Thesis, Univ. Edinburgh, Edinburgh, Scotland.
Journal of Dairy Science Vol. 76, No. 4, 1993
14 Mpofu, N. 1992. Genetic and economic evaluation of dairy cattle breeding strategies for Zimbabwe. Ph.D. Diss., Univ. Guelph, Guelph, ON, Can. 15Mpofu, N., C. Smith, W. Van Vuuren, and E. B. Burnside. 1993. Breeding strategies for genetic improvement of dairy cattle in Zimbabwe. 2. Economic evaluation. J. Dairy Sci. 76:1173. 16Pirchner, F. 1983. Population Genetics in Animal Breeding. 2nd ed. Plenum Press, New York, NY. 17 Rendel, J. M.,and A. Robertson. 1950. Estimation of genetic gain in milk yield by selection in a closed herd of dairy cattle. J. Genet. 50:l. 18Ruane. J., and R. Thompson. 1991. Comparison of simulated and theoretical results in adult MOET schemes for dairy cattle. Livest. Prod. Sci. 28:l. 19Smith. C. 1988. Breeding strategies for developing countries. World Anim. Rev. 65:2. 20Van Tassell, C. P., and L. D. Van Vleck. 1991. Estimates of genetic selection differentials and generation intervals for the four paths of selection. J. Dairy Sci. 74:1078. 21 Van Vleck, L. D. 1986. Evaluation of dairy cattle breeding programs: specialised milk production. Proc. 3rd World Congr. Genet. Appl. Livest. Prod., Lincoln, NE IX:144. 22 Woolliams, J. A,, and C. Smith. 1988. The value of indicator traits in the genetic improvement of dairy cattle. Anim. Prod. 46:333.