optimal breeding strategies for calving ease · breeding strategies for calving ease 3443 table...

GENETICS AND BREEDING

Optimal Breeding Strategies for Calving Ease

JACK C. M. DEKKERS Centre for Genetic Improvement of Livestock

Department of Animal and Poultry Science University of Guelph

Guelph, ON, Canada N1G 2W1

ABSTRACT

Calving ease is of economic importance in dairy cattle and should be considered in breeding programs. Economic values of direct and maternal calving ease were derived based on cost-benefit analysis and gene flow methodology. Marginal returns from dystocia reduction in primiparous dams were -!PI3 per phenotypic standard deviation compared with $300 for production. For an average mate, numbers of discounted expressions for sires to breed replacements were .58. .44, and .63 for direct and maternal calving ease and production. Discounted expressions for direct calving ease were higher when sires were mated to primiparous versus multiparous dams. Three alternative breeding strategies were compared: 1) separate selection of sires as mates of primiparous and multiparous cows, based on their respective optimal indexes for direct and maternal calving ease, 2) selection based on the optimal index for an average mate, and 3) selection on direct calving ease only. Strategies 2 and 3 allowed for assortative mating of sires with favorable direct calving ease to primiparous females subsequent to selection. Strategy 2 resulted in maximal economic response. Optimal standardized index weights under Cana- dian circumstances were approximately 100:9:7 for production:direct calving ease:maternal calving ease. Inclusion of calving ease traits had a minor effect (<Sa) on selection responses and effi- ciency but would provide proper guidelines to producers. Greatest benefits of genetic evaluations for calving ease were obtained through assortative mating.

Received November 17, 1993. Accepted May 23, 1994.

(Key words: dystocia, genetics, econom- ics)

Abbreviation key: CE = calving ease, DCE = direct calving ease, EBV = estimated breeding value, MCE = maternal calving ease, MGSCE = maternal grandsire calving ease.

INTRODUCTION

Dystocia is a reproductive problem of dairy cows that is of economic importance, especially for first-calf heifers. The economic costs of dystocia include loss of calf, veterinary fees, farmer labor costs, increased risk of subsequent health and fertility problems, increased culling, and reduced production (21).

Meijering (21) reviewed the biological aspects of dystocia in cattle. On a general basis, factors affecting calving ease (CE) can be separated into maternal and fetal (or direct) components. Maternal calving ease (MCE) refers to characteristics of the dam giving birth (e.g., pelvic dimensions). Direct calving ease (DCE) refers to characteristics of the calf (e.g., calf size).

Although dystocia can be reduced by proper management procedures, such as heifer rearing and feeding during gestation, selection and breeding strategies have been identified as important additional tools to reduce dystocia in the short and long term (21, 29). Many studies have identified small but significant genetic components to both DCE and MCE; heritabilities ranged from .03 to .20 (7, 13, 21, 29, 32). The DCE and MCE are genetically antagonis- tic; estimates of genetic correlations ranged between -.19 and -.63 (7, 13, 21). Most studies have found genetic correlations between CE and production traits to be close to zero (21). Significant genetic relationships between MCE and conformation traits have been found, especially conformation traits associated with pelvic shape and pelvic dimensions (4, 9, 27).

Incidence of dystocia is less for multiparous than for primiparous cows. Conflicting results

1994 J Dairy Sci 77:3441-3453 3441

3442 DEKKERS

have been reported on the genetic nature of dystocia for primiparous versus multiparous cows. Most studies found heritability of CE to be lower for multiparous than for primiparous cows (7, 30, 33), which was not entirely ex- plained by differences in incidence (33). Thompson et al. (30) and Cue and Hayes (7) found high genetic correlations between CE for primiparous and multiparous cows (.85 and 9). However, Weller et al. (33) found genetic correlations smaller than .5.

In many countries, sires are evaluated for CE based on ease of birth of their progeny, ease with which a sire’s daughters give birth, or both (3, 29, 32; K. Wade et al., 1991, unpublished data). In most cases, observations used for CE genetic evaluations consist of subjective classifications into three or four cat- egories (e.g., in Canada, unassisted, easy pull, hard pull, and surgery). Cases of abnormal presentation should be excluded from the analysis because this problem lacks a genetic component (24). A threshold model, because of its categorical nature, should be used for sire evaluation for CE. However, sire rankings that are computed based on linear and threshold models have high correlations (23, 33). Some countries evaluate sires for DCE only, but others also compute evaluations for MCE or for maternal grandsire CE (MGSCE). The MGSCE relates to a sire’s effect on ease of birth of his grandprogeny through his daughter and includes the sire’s effect for MCE plus half the sire’s effect for DCE (21, 34). In Canada, dairy sires are evaluated simultane- ously for DCE and MCE based on a reduced animal model with maternal effects, after transformation of the data using a Snell transformation (K. Wade et al., 1991, unpublished data).

Breeding strategies for CE must consider the importance of DCE and MCE for selection and mating of sires. Several studies have compared alternative breeding strategies for CE (1, 17, 22, 28). In most cases, comparisons were between preferential mating of sires to heifers based on DCE with selection on estimated breeding values (J3BV) for either DCE or MGSCE, with or without subsequent preferential mating (17,22, 28). Hanset (17) also evaluated response to selection on a weighted average of EBV for DCE and MGSCE and response to selection on an index for optimal

Journal of Dairy Science Vol. 77, No. 11. 1994

improvement of total CE (DCE plus MCE). Balcerzak et al. (1) determined the response to selection on an index of sire EBV for DCE and MCE that was derived for a breeding goal that included DCE and MCE. Balcerzak et al. (1) and Meijering (22) considered the differential economic importance of DCE and MCE or MGSCE when defining the breeding goal. This difference in economic importance results from differences in rate and timing of expression of genetic superiority of sires for DCE versus MCE (22). In both studies (1, 22), relative economic values were derived using gene flow methodology. However, Balcerzak et al. (1) did not consider differential incidence of CE for first versus later parities, and Meijering (22) assumed that dystocia was absent for multiparous females.

Few studies (2, 10, 22) have attempted to derive absolute economic values for CE or economic values of CE relative to other traits of economic importance, which is likely because of lack of accurate information on cost factors related to CE. However, economic values for CE are needed to determine the optimal emphasis on CE traits in selection relative to other traits and to quantify economic benefits from consideration of CE in breeding programs. Meijering (22) derived economic values for dystocia under economic circumstances in The Netherlands. Veterinary fees, labor, reduced production, reduced fertility, and increased culling were considered as cost components. Economic values were sensi- tive to calf price and population incidence of dystocia (22). Bekman and van Arendonk (2) used similar parameters as Meijering (22) but did not include reduced production as a cost because production was included as a trait in the breeding goal. Dematawewa (10) found that, for genetic evaluations for DCE in the US (3), an increase in sire ETA for dystocia by one standard deviation on the liability scale increased dystocia costs for resulting births by $38.79 when sires were mated to an average female in the population. Costs associated with lost production, increased days open, and lost calves were considered.

Objectives of the current study were 1) to determine the relative economic importance of DCE and MCE, considering the differential incidence of dystocia for first and later parity females, 2) to determine the economic value of

BREEDING STRATEGIES FOR CALVING EASE 3443

TABLE 1. Frequency distribution of calving ease scores by parity and sex of calf and the percentage of stillbirths within each class (in parentheses) for Canadian Holsteins.

Calving ease class Sex

Parity of calf Unassisted Easy PUU Hard pull Surgery ~ ~ ~ ~ ~ ~~

(W 1 Male 48.9 (54) 33.7 (50) 16 7 (198) .69 (34.4)

Female 55.6 (43) 32.9 (44) 11 1 (189) 40 (31.5) 22 Male 66.5 (2.4) 27 5 (2 5) 5 8 (136) 23 (34.7)

Female 71.3 (2.1) 24 8 (2 5) 3.8 (14.1) 16 (33 0)

CE traits relative to production traits, and 3) to determine optimal sire selection and mating strategies for CE based on EBV for DCE and MCE and to quantify the economic benefit of this strategy over selection on production alone. Genetic and economic parameters appropriate for Canadian Holsteins were used, but methods can be applied to other populations.

MATERIALS AND METHODS

Breeding Objective and Economlc Values

The breeding objective used in this study included additive genetic merit for production (Ap), DCE (b), and CE liability (AM=):

where vp, VDCE and VMCE are the corresponding economic values. Production was included as a dollar value: the sum of milk, fat, and protein yield, weighted by their economic values, with a genetic standard deviation of $150 Canadian (16). The CE liability represents the underlying standardized (phenotypic SD = l) continuous scale for CE, of which categorized observations (unassisted, easy pull, hard pull, and surgery) are expressions following a threshold model (15). Genetic parameters for CE used in this study were from Dwyer et al. (13) and as used for current genetic evaluations of CE in Canada: heritabilities of . I 1 and .I2 for LXE and MCE and a genetic correlation of -.27. The CE liability and production were assumed to be the same genetic traits for primiparous and multiparous cows. Genetic correlations between production and CE traits were assumed to be zero (21).

Following Brascamp (5). the economic value of trait i (vi) is the product of the marginal value of one unit of the trait (ai) and the extent to which genetic superiority for the trait is expressed over a planning horizon (DEJ. The latter is referred to as the number of discounted expressions, where "discounted" refers to discounting of individual expressions to a fixed time period (e.g., the time of selection).

Marginal Values. The marginal value of the production trait, expressed in dollars, is by definition $1 (ap = 1). Although DCE and MCE are different genetic traits, the effects of both traits are mediated and expressed through the CE. Therefore, the marginal value of one unit of CE liability (a) is the same for DCE and MCE. The e was derived using methods similar to those of Meijering (22) and Bekman and van Arendonk (2) by determining the effect of an increase in the population mean for CE liability on the proportion of calvings in each dystocia class, following a threshold model. To allow for differences in expression of dystocia and for differences in mean CE liability, four groups of calvings were distin- guished: male calves out of primiparous dams, female calves out of primiparous dams, and male and female calves out of multiparous dams. The distribution over CE classes within sex and parity groups in the Canadian Holstein population is summarized in Table 1 .

Let Ki be the average cost associated with a calving classified in class i (i = 1 to 4 for unassisted, easy pull, hard pull, and surgery); njk.the frequency of CE class i within sex and panty subclass jk (i = 1 or 2 for male and female calves; k = 1 or 2 for first and later parity dams); and tijk (i = 1, 2, or 3) the threshold on the liability scale that separates CE class i from class i + 1 within sex and

Journal of Dairy Science Vol. 77, No. 1 I , 1994

3444 DEKKERS

parity subclass jk. Let pjk be the average CE liability in subclass jk. Then, thresholds ti$ can be derived such that Pljk = F(t1jk - Pjk), where F(t) is the cumulative standard normal distribution function. Similarly, Pijk = F(t,jk - pjk) - F(G-1 j k - kjk) for 1 = 2, 3; P4jk = F(4jk - Pjk). Total costs associated with dystocia per calving in sex and parity subclass jk can be computed as

4

Cjk = PijkKi. i=l

Substitution for Pijk in this equation and taking the first derivative with regard to Pjk, the marginal value of increasing average CE liability by one unit for subclass jk can be obtained as (2, 22):

where d) represents the standard normal density function. Averaging over sexes (5050 sex ratio), the marginal value of one unit of CE liability can then be derived as

WEfi) = ( e ( 1 . 1 ) -k w,l))n

WE@) = ( e ( I . 2 ) + -0,2))n for CE of heifers and as

for CE of multiparous cows.

Average costs associated with each CE class (KJ were computed considering loss of calf, veterinary fees, extra labor costs, increased days open, and increased culling. Reduced production was not considered because inclusion of production in the breeding goal required the economic value of CE to be computed at constant production if the production trait includes the effect of CE on production. Cost components are summarized in Ta- ble 2.

Number of Discounted Expressions. The gene flow procedure developed by Brascamp (6) based on methods by Hill (18) was used to compute the number of discounted expressions when sires were selected for DCE, MCE, or production. In the gene flow model, the population was subdivided into five tiers: sires, cows born from primiparous dams, cows born

Journal of Dairy Science Vol. 77, No. 11, 1994

TABLE 2. Costs associated with factors contributing to costs per calving for calving ease (E) classes.l.2

Class for CE

Easy Hard Cost factor pull pull Surgery

Veterinary fee3 0 42.60 240.00 Farmer lab0r3.~ 3.00 5.00 15.00 Culling5 0 60.00 180.00 Days open6 6.30 37.20 67.20

'Costs were set to zero for unassisted calvings. 2costs associated with calf loss in each CE class

(including unassisted calvings) were computed by parity and sex from stillbirth frequencies and values of $75 and $130 for male and female calves from primiparous dams and of $85 and $140 for male and female calves from multiparous dams.

3Based on a survey of 17 veterinary practitioners in Canada (M. Bauman, J.C.M. Dekkers, D. Kelton, and K. Lissemore. 1993, unpublished data).

4Based on labor costs of $IO/h. 5Base.d on increases in involuntary culling by 5 and 15

percentage points for hard pull and surgery and a cost of $12 per percentage point.

6Based on results of Mangurkar et al. (19) for the effect of CE on days open and a cost of $3 per extra day open.

from multiparous dams, and calves born from primiparous and multiparous dams. Primiparous and multiparous dams were considered separately to allow for differential incidence of CE and to allow for the mating of young sires to multiparous females only. Calves were included as separate tiers because DCE is expressed in calves. Transmission of genes from parents to progeny by path of selection is presented in Table 3. The female replacement rate was set at 30%; thus, 30% of calves were born from primiparous dams. No differential selection among calves occurred with regard to parity of the dam. Therefore, 30% of herd replacements, dams of bulls, and dams of cows were born from primiparous dams. Proportional gene contributions of the pathways for dams of dams and dams of calves reflect the age distribution of cows in the population. Sires of sires and sires of cows were selected from age groups 6 through 9. In addition, 20% of multiparous cows were bred by young bulls.

Table 3 also shows the incidence vectors used for the three traits. All expressions are per cow in the population. Calving ease of


primiparous dams was used as the unit of expression for DCE and MCE, which implies that was used as the marginal value of a unit of CE expression. The difference between WE,J,) and WE(^) was accounted for through parameter c, which is the relative expression of CE for multiparous versus primiparous cows (c = CWE(~~WE,J,> Three alternative values for c were used: 0, .25, and 3.

Direct CE is expressed in calves. A relative expression of DCE for calves born from primiparous cows of 30% (Table 3) reflects that 30% of dams are primiparous cows. With MCE expressed by dams, relative expression of MCE by age of dam in Table 3 reflects the age distribution of dams of calves and a relative incidence of CE of 1 and c for primiparous and multiparous dams.

For expression of the production trait, production of an average cow in the population

was used as the unit of expression. Elements of the incidence vector in Table 3 are the product of the effect of age on production and the age distribution of cows in the population.

Expression of one unit of genetic superiority of progeny-tested sires for the trait of interest was studied by path of transmission: sires of sires, sires of dams as mates of primiparous dams, sires of dams as mates of multiparous dams, sires of calves out of primiparous dams, and sires of calves out of multiparous dams. Expressions were discounted to the year of selection of sires at a rate of 5% and accumulated over a 15-yr planning horizon. A time adjustment of .4 yr was used for discounting expressions of production to account for expression of production throughout the lactation; expression of CE is at the start of lactation.

TABLE 3. Contribution of age groups to progeny for paths of gene transmission and relative expression of direct calving ease W), maternal calving ease (MCE), and production (PROD) by age group.

~~ ~~~ ___ _ _ _ ~ _ _ ~ ~~ ~

Age group

0 yr 2 yr 3 yr 4 yr 5 yr 6 yr 7 yr 8 yr 9 yr 10 yr 11 yr

Contribution of parental age group to progeny (I) Gene transmission path

Sires of sires 40.0 35.0 20.0 5.0 Sires of dams out of

Primiparous dams 37.6 31.2 18.8 12.4 Multiparous dams 25.0 28.2 23.4 14.0 9.4

Primiparous dams 37.6 31.2 18.8 12.4 Multiparous dams 25.0 28.2 23.4 14.0 9.4

Sires of calves out of

Dams of sires' 32.8 23.6 17.2 11.5 7.1 3.6 2.1 1.5 .6 Dams of dams' 30.02 23.03 16.53 12.03 8.03 5.03 2.53 I S 3 l.03 S3 Dams of calves 30.02 23.03 16.53 12.03 8.03 5.03 2.53 1.53 l.03 S3

Relative expression of trait within age group (%)

Trait and population tier DCE4 Calves out of 30.0 Primiparous dams 70.0~5 Multiparous dams

MCE4 Dams' 30.0 23.1~5 16.4~5 12.0~5 8 .0~5 5 . 1 ~ 5 1.8~5 2 .3~5 l.lc5 .6c5 PROD6 Dams' 25.6 22.4 17.5 13.4 9.1 5.7 2.9 1.7 1.1 .6

*In each age group, 30% of animals were out of primiparous dams and 70% out of multiparous dams Tontribution to progeny out of primiparous dams. Tontribution to progeny out of multiparous dams. 4Expression relative to primiparous dams. 5Incidence of dystocia of multiparous dams relative to that of primiparous dams. 6Expression relative to a female of average age.


3446 DEKKERS

Sire Genetic Evaluations for CE

In Canada, EBV of sires for DCE and MCE are estimated using an animal model with maternal effects (K. Wade et al., 1991, unpublished data). Under such a model, the main sources of information for evaluation of progeny-tested sires are ease of birth of a sire’s calves and CE records on a sire’s daughters when used as dams. Here, sire EBV for both DCE and MCE were approximated using selection index theory (8) as a linear function of average CE of a sire’s progeny and grandprogeny. Variances and covariances similar to those used by Hanset (17) and Balcerzak et al. (1) were used to derive the index equations to estimate EBV. However, unlike methods by Hanset (17) and Balcerzak et al. (l), daughters giving birth to a sire’s grandprogeny were assumed to be a subset of a sire’s progeny that contributed to the average CE of his progeny. The categorical nature of CE observations was ignored. Sire EBV for production were approximated based on the average performance of daughters calving.

Strategies for Sire Selection

Because of the assumed zero genetic and phenotypic correlations between production and CE traits (21) and because EBV for DCE and MCE are obtained from common sources of information (CE of progeny and grandprogeny), similar to multiple trait evaluation, the optimal index 0 for selection of sires for breeding goal T is obtained by simply r$plac- ing true breeding values (A,) by EBV (A,) in T, which can be shown using selection index theory (8):

I = ~ p A p + v D C E A ~ E + v M ~ A M ~ .

Optimal relative emphasis on DCE versus MCE in the index is equal to the ratio of discounted expressions ( D E ~ ~ / D E M ~ ) , because marginal returns of a single expression (WE) are the same for DCE and MCE and v = WE x DE. Response to one standard deviation of selection on I was obtained for T, traits included in T, and correlated traits, using selection index theory (8).

Because of differences in numbers of discounted expressions for DCE and MCE, de-

pending on whether the sire is mated to primiparous or multiparous cows, the optimal index for selection of sires of herd replacements differs among mates. In addition, assortative mating can be used. To explore these possibilities, responses to three strategies for selection of sires of dams were investigated:

1) dual index, select sires to be mated to primiparous and multiparous cows separately based on their respective optimal indexes;

2) single index, select sires based on an index that is optimum for an average mate; and

3) direct only, select sires on EBV for DCE only.

To illustrate differences clearly, selection strategies were compared for a breeding goal in which the economic value of production was set equal to zero (vp = 0) and for no incidence of dystocia for multiparous cows (c = 0).

Impact of Assortative Mating

Strategies that use the same index to select sires for all mates (single index and direct only strategies) allow for assortative mating of sires subsequent to selection on I, by mating sires with favorable EBV for DCE to primiparous cows. If it is assumed that a fraction p of selected bulls are selected on EBV for DCE to be mated to primiparous cows, then, ignoring the effects of selection on (co-)variances, which is valid if emphasis on CE in the initial selection is small, the effect of assortative mating on economic returns per cow mated is

RAM = WE (~~DEDcE(sc~N - 11-p DEDcE(sc(C))) ~ A D c E ) ~

where ip. and il-, are selection intensities, U(AXE) IS the standard deviation of sire EBV

are the number of discounted expressions for DCE from the paths of sires to calves out of primiparous and multiparous dams (22).

for DCE, and DEDCE(SC(h)) and DEDCE(SC(c))

RESULTS

Table 1 shows the frequency of CE scores by parity and sex for the Canadian Holstein



population and the percentage of calves that are stillborn (death at birth or within 24 h after birth) in each category. Frequencies of hard pull and surgery were approximately three times as high for primiparous as for multiparous dams. However, frequencies of dystocia cannot be ignored for later parities. The frequency of stillbirths was substantial within the hard pull and surgery classes and was little affected by sex or parity. However, the frequency of stillbirths for calvings classified as unassisted or easy pull was twice as high for primiparous as for multiparous cows. Results correspond with previous studies of this and other Holstein populations (7, 20, 30).

Economic Values

Marginal Values. Total costs per case by CE class, parity of dam, and sex of calf are in Table 4. Withm sex and CE class, costs were little affected by parity. In fact, stillbirth (Table 1) was the only cause of a parity effect on costs within a class. Dematawewa (10) found a strong increase in costs within CE class with parity, whch was due to a larger impact of dystocia on production for later parity cows (11).

Averaging over sexes resulted in marginal values of close to $43 for primiparous dams (-9 and of $20 for multiparous dams

(Table 4). This result indicates that, although the frequency of dystocia is three times as high for primiparous as for mul-

tiparous cows, dystocia is, in economic terms, only twice as important for first parity dams. The difference is because of the nonlinear relationship between frequency of dystocia and dystocia liability. Dematawewa (10) also found the difference in dystocia between parities to be substantially smaller in economic terms than on the basis of incidence, which was mainly because of a greater effect of dystocia on production losses for multiparous than for

A reduction in the frequency of hard pull cases was the main contributor to the marginal value of CE liability (Table 4). A 50% reduction in veterinary fees, labor, culling, days open, and calf losses associated with dystocia reduced marginal values for CE by approximately 14, 2.5, 17, 13, and 13%. Removal of all costs of culling from the economic value of CE, which would be appropriate if involuntary culling or herd life is included as a trait in the breeding goal, reduced the marginal value by over 30% (Table 4).

Number of Discounted Expressions. Figure 1 shows the pattern of expression of one unit of genetic superiority for production, DCE, and MCE of sires selected in yr 0 as sires of replacements. All expressions are per success- ful insemination or, equivalently, per dam of the corresponding parity in the population. For dystocia traits, expressions are in heifer equivalents and assume that expression of dystocia for cows is 50% of that for heifers (c = S). For production, expressions are in terms of production for a cow of average age.

primiparous cows.

TABLE 4. Total costs per case relative to unassisted calvings by calving ease (02) class, parity, and sex of calf, and the marginal value of calving ease liability per phenotypic standard deviation.

Class CE

Parity 1 Parity 22

Male calf

Female Male calf calf

Female calf

(9 Easy pull 9.00 9.43 9.39 9.86 Hard pull 155.60 163.39 154.32 161.60 surgery 523.95 537.56 529.66 545.46

($ per phenotypic SD) Marginal value 48.27 37.95 23.05 17.45

43.1 1 20.25' (27.74)'~~ (1 3.57)'32

'Averaged over male and female calves, by parity *Excluding costs from premature culling.

Journal of Dairy Science Vol. 77, No. 1 1 , 1994

3448 DEKKERS

TABLE 5 . Number of discounted expressions1 of sire genetic merit for production, direct calving ease (DCE), maternal calving ease (MCE), and DCE relative to MCE (DCUMCE) over a planning horizon of 15 yr and at an interest rate of 5%. by path of transmission, and parity of mate.

Parity of mate Expression for Path of multiparous dams Trait transmission2 1 22 Average

(% of primiparous dams) Production3 SS .17 . I 7 .17 SD .63 .63 .63

25 DCE ss .I5 .I5 .15 SD .20 .20 .20 sc .48 .I2 .23 SD + SC .68 .32 .43

MCE3 ss .10 .10 . l O SD .33 .33 .33

DCUMCE ss 1.51 1.51 1.51 SD + SC 2.04 .95 1.28

W E ss .20 .20 .20 SD .21 .27 .27 sc .48 .24 .3 1 SD + SC .75 .51 .58

MCE3 ss . I 3 .13 .I3 SD .44 .44 .44

DCUMCE ss 1.60 1.60 1.60 SD + SC 1.69 1.15 1.31

‘Expressions are per female of the corresponding parity in the population and expressed on a heifer equivalents basis for dystocia traits and on the basis of a cow of average age for production.

2SS = Sires of sons, SD = sires of dams, and SC = sires of calves. 3For production and MCE, the path SC contributes zero discounted expression.

50

Figure 1 clearly illustrates that a sire’s genes are expressed to different degrees and at different rates for the three traits, thus making use of discounted expressions necessary when economic values are computed. Parity of the sire’s mate is important only for DCE and affects only the degree of expression in yr 1, at the time of birth of the progeny.

Table 5 shows the number of discounted expressions of one unit of genetic merit of sires for production, DCE, and MCE. Numbers of discounted expressions are per cow of the corresponding parity in the population. For dystocia traits, results are presented for expression of dystocia for multiparous dams at 25 and 50% relative to primiparous dams (c = .25 and c = S). A level of 50% corresponds to the relative marginal value of dystocia for primiparous versus multiparous dams (Table 4).

Parity of the mate had an effect only on the number of discounted expressions contributed by the sires of calves path for DCE (Table 5). For DCE, sires of calves was the most important path to consider for the mating of a primiparous female. All paths (sires of sires,

sires of dams, and sires of calves) were of similar importance for the mating of a multiparous or average cow, especially for a 50%

1

0 -

YEAR

Figure 1. Expression of one unit of genetic superiority for direct calving ease (DCE depending on parity of the mate), maternal calving ease (MCE), and production (P) of sires that are selected as sires of herd replacements in year zero. Incidence of dystocia for multiparous dams is 50% of the incidence for primiparous dams. Expressions are in heifer equivalents for DCE and MCE and relative to an average cow for production.

Journal of Dairy Science Vol. 77, No. 1 1 , 1994


incidence rate for multiparous cows (Table 5). The ratio of number of discounted expressions for DCE and MCE (DCEMCE in Table 5 ) quantifies the relative importance of the two traits in the breeding goal and selection index. For selection of sires of sons, DCE was 50 to 60% more important than MCE. The relative importance of DCE may be limited for this path by the strategy of mating young bulls only to multiparous cows.

For selection of sires of herd replacements (sues of dams plus sires of calves) to be mated to heifers, DCE was 100 and 70% more important than MCE for relative incidences of dystocia for multiparous cows of 25 and 50%. For selection of sires as mates of multiparous cows, DCE was 5% less and 15% more important than MCE (Table 5). Averaged over mates, DCE was approximately 30% more important than MCE, regardless of incidence of dystocia for multiparous cows. Balcerzak et al. (1) found DCE to be almost three times as important as MCE. The difference relates to the shorter time horizon (10 yr) and higher interest rate (10%) used in that study and the assump- tion that 1 of 6 females mated produced a herd replacement, compared with 1 of 3.3 females in the current study. The latter is consistent with a 30% annual replacement rate. Meijering (22) found that the number of discounted expressions for DCE was 32% higher than for MGSCE. However, comparisons with the current study are hampered by differences in length of the planning horizon and by alternative definitions of CE traits.

For production, the sires of dams path was over three times as important as the sires of sires path (Table 5). The lower relative importance of the sires of sires path relative to that of other studies (5, 22) is the result of the shorter planning horizon considered here (15 versus 25 or 30 yr). Another difference is that production was assumed to be the same trait across lactations; previous studies (5 , 22) assumed a .8 genetic correlation between production for first and later lactations.

Comparison of Strategies for Sire Selection

Table 6 shows results for the comparison of the three strategies for selection of sires of herd replacements. To illustrate differences, results are shown only for a situation with no

TABLE 6 . Response in direct (DCE) and maternal (MCE) calving ease of daughters and in number of discounted expressions to one standard deviation of selection of sires for calving ease based on t h e alternative selection strategies.'

~ ~ ~ ~~

Single Dual Direct index index only

Index weight DCE I .22 2.68*/.583 I MCE 1 1 0

SD) DCE .10 .08 .14 MCE .06 .08 -.04 DCE + MCE .16 .I5 .10

Parity 1 ,155 ,168 ,158 Parity 22 ,055 .062 ,022 Mean ,085 ,094 .063

Response (phenotypic

Discounted expressions4

lSingle index = Selection on index that is optimal for average cow, dual index = selection of sires by parity of mate (first versus later) based on the respective optimal index, and direct only = selection of sires based on direct calving ease only. Estimated breeding values for DCE and MCE are based on records on 100 calves born and 50 daughters calving. Indexes and responses were derived for a situation with dystocia among heifers only.

zlndex for selection of mates of first parity dams. slndex for selection of mates of multiparous dams. 4Number of discounted expressions of heifer calving

ease equivalents per female mated based on a planning horizon of 15 yr and a discount rate of 5%.

incidence of dystocia for multiparous cows (c = 0) and no selection on production (vp = 0).

As expected, selection on only DCE resulted in the greatest response in DCE. How- ever, a negative correlated response in MCE resulted because of a negative genetic correlation ("able 6). Improvement in total CE (DCE plus MCE) was over 30% lower than for strategies that considered both DCE and MCE. The single and dual index strategies resulted in similar responses in total CE liability. How- ever, with a single index, the majority of improvement resulted from improvement of DCE; DCE and MCE improved equally with dual index selection. Based on number of discounted expressions, the dual index was eco- nomically superior by 10% over the single index strategy.

Although the dual index approach allows for customization of selection indexes to parity

Journal of Dairy Science Vol. 77, No. 11 , 1994

3450 DEKKERS

of the mate, single index approaches (including selection on only DCE) allow for selective mating of sires to primiparous and multiparous cows subsequent to selection. When 30% of selected sires are selected to be mated to primiparous cows based on EBV for N E , dystocia liability for the resulting births would be reduced by .16 phenotypic SD compared with that for random mating. This translates into .158 discounted expressions and, with a 10% incidence of dystocia among heifers under random mating, into a reduction of dystocia incidence by 2.6%. However, if dystocia is present also for multiparous cows, assortative

and .024 if relative incidence for multiparous cows is 0, 25, and 50%. These figures can be compared directly with a difference in genetic response of .009 discounted expressions between the single and dual index approach (Ta- ble 5). which indicates that assortative mating would more than compensate for lower genetic response from selection on a single index. Superiority of the single index with subsequent mating would be even larger when other traits are considered in selecting sires, because benefits from assortative mating would be little affwted by other traits.

mating would result -in a slight increase in dystocia for multiParous cows by s o l 7 and .034 discounted expressions, if incidence for

Sire Selection for CE in Combination with Production

multiparous cows 25 and 50% of that for Table 7 shows optimal weights for a selec- heifers. Averaged over parities, assortative tion index and responses to selection when mating would result in increases in discounted both CE and production traits are considered in expressions per female mated by .047, .035, the breeding goal and the selection index,

TABLE 7. Relative economic values and index weights for direct and maternal (MCE) calving ease when sires are selected based on an index of estimated breeding values for DCE, MCE, and production dollar value,l responses in calving ease and production of daughters and in discounted returns to one standard deviation of sire selection, and responses to assortative mating of sins to primiparous females based on DCE for different economic values for calving ease.

Economic value of calving ease2

25% Dystocia3 50% Dystocia

Item

Standardized' economic value and index weight relative to production DCE .093 .067 ,044 .031 ,127 ,091 ,059 .042 MCE ,076 .OS5 .036 ,025 .lo1 ,072 ,072 .034

sire selection

DCE (up x 100) 1.02 .74 .48 .34 1.39 1.00 6 5 .47 MCE (u x 100) .54 .39 .25 .I8 .70 .51 .33 .24 DCE + L C E (up x 100) Discounted returns, %a .46 .23 .10 .OS 3 3 .43 . I8 .09

Returns from assortative mating 2.12 1.52 99 .71 1.41 1.01 .66 .47

60 $/up 43 $/up 28 $/up 20 $/ap 60 $/up 43 $/up 28 gap 20 Vup

Response to one standard deviation of

Production 465 -.45 -.23 -.IO -.OS -.82 -.42 -.18 -.09

($ per female)

'Sire estimated breeding values based on records on 100 calves born and 50 daughters calving and lactating. 2Dollars per phenotypic standard deviation (up). 3Incidence of dystocia for multiparous relative to primiparous dams. 4Standardized by genetic standard deviation. SResponse in dollar value of production for daughters as a percentage of response to selection on production only

6Discounted returns as a percentage of response to selection on production only ($81.51). based on a planning horizon ($64.49).

of I5 yr and an interest rate of 5%.


BREEDING STRATEGIES FOR CALVING EASE 345 1

weighted by their economic values. Results are shown for a range of marginal values for CE, including values of $43 and $28 per standard deviation of CE liability, as obtained in the current study (Table 4). Results are shown only for selection of sires of replacements on a single CE index that is optimal for an average mate. The effect of subsequent assortative mating is presented also.

For prediction of response and comparison of alternative strategies, relative rather than absolute economic values of traits are important. Therefore, relative responses in Table 7 hold for any situation with corresponding relative marginal values on the basis of genetic standard deviations.

Results in Table 7 show that, depending on the economic value of CE, the sum of standardized index weights on DCE and MCE was less than 17% of that on production in most cases. Selection on such an index would result in a less than .5% reduction in response for production, compared with selection on production only, and in small genetic improve- ments in CE; response in DCE would be almost twice as large as response in MCE. Con- sideration of CE traits for selection would increase discounted returns from genetic improvement by less than .5% in the most likely situations. With 20% of available AI bulls selected as sires of herd replacements, a .5% increase in returns would represent $.42 per cow mated, or $420,000 for a population of one million cows. The greatest benefits from including CE traits in a breeding strategy origi- nate from assortative mating based on DCE, which would net from $.70 to $1.20 per female mated under the most likely scenarios (Table 7).

DISCUSSION

Based on results from this study, the optimal breeding strategy for CE is selection of sires on an index that includes EBV for both DCE and MCE, along with other traits of economic importance, and subsequent assortative mating of sires with favorable EBV for DCE to primiparous cows. Weights on EBV for DCE and MCE in a total merit index should be based on their economic impact when selected sires are mated to an average cow. Under current Canadian circumstances,

standardized weights in an index that includes sire EBV for production value, DCE, and MCE should be approximately 100:9:7 (Table 7; 50% incidence for multiparous cows and marginal value for CE of $43 per phenotypic SD). When EBV for herd life or traits related to herd life are also included in the index, standardized weights on DCE and MCE should be reduced to up to 6 and 4.5 (Table 7), depending on the genetic relationships between CE and herd life (associated) traits.

Economic parameters for CE are not well known, which could affect conclusions from this study. One aspect of dystocia that was not accounted for when economic value was computed is the impact of dystocia on risk of subsequent diseases, which has been well es- tablished (12, 14, 25), but results are not consistent. Costs associated with increased risk of subsequent disease (and culling) can be estimated by path analysis methods, as illustrated by Oltenacu and Lindhe (26), and should be accounted for in the economic value of CE if they are not accounted for in other traits in the breeding goal.

Some studies have quantified the significant negative effects of dystocia on production during the subsequent lactation of the cow giving birth (11, 21, 31). but others did not find a large effect (19. 21). Djemali et al. (11) found that the effect of dystocia on production increased with parity. Lost production was not included in the present study as a cost of dystocia, because production was included as a trait in the breeding goal. However, depending on the definition of the production trait, part of these costs may need to be accounted for in the economic value of CE.

Cost of lost production and premature culling should be accounted for when economic benefits from assortative mating are computed, which was not done in the current study. As a result, benefits from assortative mating might have been underestimated. However, if the negative effect of dystocia on production increases with parity (1 1). the degree of underes- timation would be small.

The main benefit of genetic evaluation of sires for CE is derived from use of EBV of DCE for assortative mating, which agrees with results of Philipsson (28) and Meijering (22). Inclusion of EBV for DCE and MCE in a total merit index does not result in significant


3452 DEKKERS

genetic improvement of CE or increases in profitability. However, such a selection strategy can nevertheless be recommended for several reasons. Most important, inclusion of CE traits in the total merit index gives proper guidelines to producers regarding utilization of EBV for CE and avoids current practices of independent culling of sires for CE. Inclusion of CE traits in a total merit index with weights based on their relative economic values does not result in significant reductions in genetic improvement of other traits of economic importance. Independent culling on CE could have a much greater negative impact on genetic improvement of other economic traits. In addition, inclusion of CE traits in a total merit index avoids deterioration of CE and ensures long-term reproductive viability. Fi- nally, the aspect of animal welfare, which is becoming increasingly important for proper strategies for genetic improvement, should be considered.

ACKNOWLEDGMENTS

Financial support for this study from the Canadian Association of Animal Breeders, the Holstein Association of Canada, and the Natu- ral Sciences and Engineering Research Council of Canada is gratefully acknowledged.

REFERENCES

1 Balcemk, K. M., A. E. Freeman, and R. L. Willham. 1989. Selection for direct and maternal genetic effects for dystocia in Holsteins. J. Dairy Sci. 721273.

2 Bekman, H., and J.A.M. van Arendonk. 1993. Deriva- tion of economic values for veal, beef and milk production traits using profit equations. Livest. Prod. Sci. 34:35.

3 Berger, P. J. 1994. Genetic prediction for calving ease in the United States: data, models. and use by the dairy industry. J. Dairy Sci. 77:1146.

4 Boldman, K. G., and T. R. Famula. 1985. Association of sue dystocia transmitting ability with progeny linear type traits in Holsteins. J. Dairy Sci. 68:2052.

5 Brascamp. E. W. 1973. Model calculations concerning optimization of A.I. breeding with cattle. 1. The economic value of genetic improvement of milk yield. 2. Tierz. Zuechtungsbiol. 9O:l.

6 Brascamp, E. W. 1978. Methods on economic optimization of animal breeding plans. Rep. B-134, Res. Inst. Anim. Prod. “Schoonoord”, Zeist, The Nether- lands.

7 Cue, R. J., and J. F. Hayes. 1985. Correlations of various direct and maternal effects for calving ease. J. Dairy Sci. 68:374.

SCunningharn, E. P. 1969. Animal Breeding Theory. Landbruksbok-handelen, Vollebekk, Denmark.

9Dadati. F. B., B. W. Kennedy, and E. B. Burnside. 1985. Relationships between conformation and reproduction in Holstein cows: type and calving performance. J. Dairy Sci. 68:2639.

10 Dematawewa, C.M.B. 1993. Cost and risk evaluation and optimization of young sire use in terms of dystocia in dairy breeding. M.S. Thesis, Iowa State Univ., Ames.

11 qemali, M., A. E. Freeman, and P. J. Berger. 1987. Reporting of dystocia scores and effects of dystocia on production, days open and days dry from Dary Herd Improvement data. J. Dairy Sci. 702127.

12Doh00, I. R., and S. W. Maain. 1984. Disease, production. and culling in Holstein-Friesian cows. 111. Disease and production as determinants of disease. Prev. Vet. Med. 2:671.

13 Dwyer, D. J., L. R. Schaeffer, and B. W. Kennedy. 1986. Bias due to corrective matings in sire evaluations for calving ease. J. Dairy Sci. 69:794.

14Erb. H. N., R. D. Smith, P. A. Oltenacu, C. L. Guard, R. B. Hillman, P. A. Powers, M. C. Smith, and M. E. White. 1985. Path model of reproductive disorders and performance, milk fever, mastitis, milk yield, and culling in Holstein cows. J. Diury Sci. 68:3337.

15Falconer, D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. John Wiley & Sons, Inc., New York, NY.

16Gibson, J. P., N. Graham, and E. B. Burnside. 1992. Selection indexes for production traits of Canadian dairy sires. Can. J. Anim. Sci. 72:474.

17 Hanset, R. 1981. Selection problems when antagonis- tic effects exist between production characteristics and calving difficulties. Livest. Prod. Sci. 8:291.

18Hill. W. G. 1974. Prediction and evaluation of response to selection with overlapping generations. Anim. prod. 18:117.

19Mangurkar. R. R., J. F. Hayes, and J. E. Moxley. 1984. Effects of calving ease-calf survival on production and reproduction in Holsteins. J. Dairy Sci. 67: 1496.

20Martinez. M. L., A. E. Freeman, and P. J. Berger. 1983. Genetic relationship between calf livability and calving difficulty of Holsteins. J. Dairy Sci. 66:1494.

21 Meijering, A. 1984. Dystocia and stillbirth in cattle- review of causes, relations, and implications. Livest. Prod. Sci. 11:143.

22 Meijering, A. 1986. Dystocia in dairy cattle breeding with special attention to sire evaluation for categorical traits. Ph.D. Diss., Wageningen Agric. Univ.. Wageningen, The Netherlands.

23 Meijering, A., and D. Gianola. 1985. Linear versus non-linear methods of sue evaluation for categorical traits: a simulation study. Genet. Sel. Evol. 17:115.

24 Meijering, A,, and P. Van Eldik. 1981. Proefproject geboorteregistratie. Rep. B-165, Res. Inst. Anim. Prod. “Schoonoord”, Zeist, The Netherlands.

25Oltenacu, P. A., A. Frick, and B. Lindhe. 1990. Epidemiological study of several clinical diseases, reproductive performance, and culling in primiparous Swedish cattle. Prev. Vet. Med. 959.

26 Oltenacu, P. A., and B. Lindhe. 1989. Risk of diseases and projected financial losses. A modelling approach.



Eur. Assoc. Anim. Prod. Annu. Mtg.. Dublin, Ireland. 27 Philipsson, J. 1976. Studies on calving difficulty, still-

birth, and associated factors in Swedish canle breeds. IV. Relationships between calving performance, precalving body measurements and size of pelvic opening in Friesian heifers. Acta Agric. Scand. 26: 221.

28 Philipsson, J. 1979. Breeding for calving performance. Page 189 in Calving Problems and Early Viability of the Calf. B. Hoffmann, I. L. Mason, and J. Schmidt, ed. Martinus Nijhoff, The Hague, The Netherlands.

29 Philipsson, J.. J. L. Foulley, J. Lederer, T. Libonussen, and A. Osinga. 1979. Sire evaluation standards and breeding strategies for limiting dystocia and stillbirth. Livest. Prod. Sci. 6:111.

30Thompson. J. R., A. E. Freeman, and P. J. Berger.

1981. Age of dam and maternal effects for dystocia in Holsteins. J. Dairy Sci. 64:1603.

31 Thompson, J. R., E. J. Pollak, and C. L. Pelissier. 1983. Interrelationships of paxturition problems, production of subsequent lactation, reproduction, and age at first calving. J. Dairy Sci. 66:1119.

32 Weller, J. I., and D. Gianola. 1989. Models for genetic analysis of dystocia and calf mortality. J. Dairy Sci. 72:2633.

33 Weller, J. I., 1. Misztal, and D. Gianola. 1988. Genetic analysis of dystocia and calf mortality in Israeli- Holsteins by threshold and linear models. J. Dairy Sci. 71:2491.

34 Willham, R. L. 1972. The role of maternal effects in animal breeding. 111. Biometrical aspects of maternal effects in animals. J. Anim. Sci. 36:1288.


optimal breeding strategies for calving ease · breeding strategies for calving ease 3443 table...

Documents