| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ANIMAL GENETICS |
USDA, ARS, US Meat Animal Research Center, Clay Center, Nebraska 68933-0166
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
Key Words: birth weight • calving difficulty • cattle • estimated breeding value • genetic trend • selection
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
reported that first-calf heifers had nearly 50% more calving difficulty than mature cows. Brinks et al. (1973) and Laster et al. (1973) found that dystocia decreased the survival of calves and caused a reduction in probability of subsequent conception. Rogers et al. (2004) found that cows experiencing dystocia were at a 58% greater risk of subsequent culling.
Response to simple selection for heifer calving ease would be slow, because the trait is observed only in calves born to heifers and heritability is low to moderate (Koots et al., 1994; Bennett and Gregory, 2001a). However, heifer calving ease is highly correlated with birth weight and moderately correlated with other weights measured on most animals in a herd (Bennett and Gregory, 2001a). Selection using multiple-trait BLUP accounting for both correlated traits and measurements on relatives should increase the selection response well beyond simple selection for heifer calving ease.
Selection only for calving ease would decrease growth, especially when calving ease breeding value predictions incorporate correlated weight traits. Selection indexes and schemes that mitigate correlated changes in growth have been proposed (Dickerson et al., 1974; Foulley, 1976) and experimentally evaluated (Arnold et al., 1990; MacNeil et al., 1998; MacNeil, 2003).
The objective of this research was to estimate changes in breeding value in beef cattle experimentally selected for improved calving ease while maintaining or increasing growth by using EBV predicted from multiple-trait BLUP. Results will determine the genetic response realized through multiple-trait selection on moderately to highly antagonistic traits.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
Populations
The Germ Plasm Utilization (GPU) experiment at the US Meat Animal Research Center, Clay Center, NE (Gregory et al., 1991a), compared 9 purebred and 3 composite populations from 1978 through 1992 and was then terminated. Four purebred (Angus, Charolais, Gelbvieh, and Hereford) and 3 composite (MARC I, MARC II, and MARC III) populations were transferred to this experiment. Some animals from other US Meat Animal Research Center sources were also added to increase population sizes.
Two lines (select and control) were formed in each of the 7 populations. The 2 lines were intended to contrast breeding value selection for decreased 2-yr-old calving difficulty with no change in calving difficulty. Details of selection differed by population and are described in a subsequent section. Herd size was approximately 150 calving females (including 45 two-year-old heifers) for each select line and 42 calving females (including 12 two-year-old heifers) for each control line. The first selected parents were mated in 1992 and 6,926 select and 2,043 control calves were born from 1993 through 1999.
Purebred and composite populations both used bulls and semen from bulls born within these lines. In addition, the purebred and composite populations both used GPU bulls that were alive in 1992 and semen collected through 1992. When the experiment started, 29 Angus, 24 Charolais, 27 Gelbvieh, 19 Hereford, 84 MARC I, 66 MARC II, and 37 MARC III sires used in GPU had more that 15 units of semen remaining in frozen storage. Because bulls born in GPU before 1993 were not assigned to lines, these bulls were used based on their EBV. The EBV of 30 GPU bulls used in a line changed enough during the experiment that they were subsequently used in the other line. When this happened, progeny remained in the line designated at the time of mating.
The purebred populations used some sires introduced from the industry. Purebred bulls introduced from the industry were assigned to control or select lines. In Herefords, a group of bulls obtained from a single source before the experiment began, and not clearly compatible with either control or select criteria, were used in both lines to a limited extent based on within-herd EBV.
Approximately 15 select and 6 control AI and natural service sires were bred to females each year, minimizing matings of close relatives. Select line AI sires were used until semen supplies were depleted or sires with better EBV became available. Select line natural service sires were retained until younger sires with better EBV were available or they became unsound. Control line AI and natural service sires were replaced more quickly and used on fewer females than select line sires in an effort to increase effective numbers. Total numbers of unique bulls used were 351 (53 industry) in select lines and 235 (35 industry) in control lines.
Before selection in 1992, 109 Angus, 155 Charolais, 102 Gelbvieh, 116 Hereford, 233 MARC I, 221 MARC II, and 240 MARC III cows were assigned to the experiment. Before selection in 1992 and in 1993, 276 Angus, 114 Charolais, 78 Gelbvieh, 174 Hereford, 202 MARC I, 200 MARC II, and 185 MARC III yearling heifers born in 1991 or 1992 were assigned to the experiment. In populations with an excess of females assigned, females were first assigned to control lines based on control line criteria and sire. The remaining cows were assigned to select lines, and the desired number was retained based on select line criteria and sire. In the Gelbvieh population, females were assigned to control and select lines at random within sire. Some cows (388) produced progeny born in both lines during the first 3 yr. Cows were changed to a different line based on updated EBV and a desire to maintain high effective numbers of parents in the smaller control lines. When cows’ lines were changed, their calves remained in the lines designated at breeding. Cows remained within the lines designated for the 1994 breeding season for subsequent years. Cows in select and control lines were culled based on a negative pregnancy diagnosis, on EBV selection criteria, and on health and temperament.
Management
Yearling heifers were bred AI for approximately 21 d, followed by natural service bulls in individual pastures for approximately 42 d. Only yearling bulls were bred by natural service to yearling heifers. Approximately 3 wk after beginning AI mating of heifers, cows 2 yr old and older were bred AI for 21 d, followed by natural service bulls 2 yr old and older in individual pastures for approximately 42 d. Sires used for AI mating were used for both heifers and older cows. In 1996 and 1997, 12 to 13% of cows (1 select line sire mated to 24 cows and 1 control line sire mated to 7 cows in each population) were bred only by natural service bulls in single-sire pastures for the entire 9-wk breeding season.
Except when cows and their calves were in single-sire pastures, select and control line females from the same population were mixed within contemporary calving groups. Average calving dates were March 15 for 2-yr-old heifers and April 6 for cows 3 yr old and older. Heifers calving as 2 yr olds averaged 712 d of age (90% of heifers calved between 674 and 753 d of age). A mixture of corn silage and alfalfa haylage, along with alfalfa and grass hay, was fed to 2-yr-old females starting approximately 2 mo before calving and continuing until adequate pastures were available, usually in middle to late April. Older females were fed limited quantities of corn silage and alfalfa haylage to meet nutrient requirements from November until mid- to late April.
The average weaning date was October 14. Average weaning ages of calves were 212 d (yearly averages from 200 to 225 d) from 2-yr-old heifers and 190 d (yearly averages from 177 to 203 d) from cows 3 yr old and older. After an initial adjustment feeding period of approximately 42 d, females were fed diets composed of corn silage, alfalfa haylage, and a protein-mineral-vitamin supplement in various proportions (approximately 2.2 to 2.3 Mcal of ME/kg of DM) and lengths of time, depending on weather conditions and weight gains, until they were placed on improved cool-season grass pasture from mid- to late April. Hereford heifer calves born in 1998 and 1999 were managed differently and were fed for slaughter after weaning. After a 42-d adjustment period after weaning, males were fed a diet composed of corn silage, rolled corn, and a protein-mineral-vitamin supplement (approximately 2.7 Mcal of ME/kg of DM).
Calves were weighed at birth, at weaning, and at 148 d after weaning (yearly averages from 140 to 157 d). Yearling heifers were weighed when they were palpated for pregnancy. Subsequently, females were weighed, measured for height, and scored for condition before calving, at the start of breeding season, and when they were palpated for pregnancy (3 to 21 d after weaning) each year.
Calving difficulty was subjectively evaluated by field personnel trained each year for accuracy and consistency of calving difficulty scores. The following descriptive scores were used: 1 = no difficulty, 2 = little difficulty by hand, 3 = little difficulty with a calf jack, 4 = slight difficulty with a calf jack, 5 = moderate difficulty with a calf jack, 6 = major difficulty with a calf jack, 7 = cesarean birth, and 8 = abnormal presentation.
Selection
Data. Four traits were used in the EBV analysis: calving difficulty score for 2-yr-old heifers, birth weight, weaning weight adjusted to 200 d, and postweaning gain adjusted to 168 d. Actual weaning weights were adjusted to 200 d, assuming linear growth from birth to weaning. Postweaning gain was adjusted to 168 d, assuming linear growth from weaning to yearling weight measurement. Heifer calving difficulty scores were set to missing values for all calves born to cows 3 yr old and older and for any calf scored an 8. All twin calf data and weaning weights and postweaning gains of fostered calves were set to missing values for analysis.
EBV.
Single-trait analyses for EBV were used to make selections in 1992 and 1993. Subsequently, EBV were estimated from a 4-trait animal model by using MTDFREML (Boldman et al., 1995). Direct genetic effects were modeled for all 4 traits. Maternal genetic effects were modeled for heifer calving difficulty score, birth weight, and weaning weight. Permanent environmental effects attributable to dams were modeled for birth and weaning weights. Estimated (co)variances (Bennett and Gregory, 2001a) were used for random effects. Fixed contemporary groups were defined by combinations of year of birth, sex, and management group. Age of dam (2, 3, 4, and 5 yr or more) was fitted with linear and quadratic regression coefficients. In purebred populations, genetic groups (USMARC source, select industry source, and control industry source) were used (Westell et al., 1988). In Herefords, there was an additional grouping for the bulls obtained from a single source before the beginning of the experiment.
Each population was analyzed independently, including data from GPU and other contributing experimental sources beginning in 1978. Each population was analyzed ignoring line (select or control).
Selection Procedures.
The selection goals were the creation of 2 lines for each population with similar growth to yearling age, but with improved heifer calving ease in one line. The lines were created by using an ad hoc multistage selection procedure based on multi-trait EBV. The objectives of the ad hoc procedure were similar to those of the restricted and desired gain indexes (Brascamp, 1984), but it was unclear how these indexes could be applied directly in this complex situation with different sire sources, multiple generations, and multistage selection using EBV.
Every year, EBV were analyzed 3 times. Each time, target EBV for maternal genetic weaning weight and direct genetic yearling weight were identified for each population. The same EBV targets were used in both selection and control lines within populations. In select lines, animals were selected with the lowest EBV for calving difficulty score and with EBV for maternal genetic weaning weight and direct genetic yearling weight within an acceptable range around their target EBV. In control lines, target EBV for direct genetic birth weight were identified in each population that would result in birth weight changes proportionally similar to yearling weight changes. The proportion was determined from phenotypic averages for birth and yearling weights. Animals in control lines were selected from those with EBV within an acceptable range around targets for direct genetic birth weight, maternal genetic weaning weight, and direct genetic yearling weight. Additional constraints on numbers of half-sibs selected were used to maintain a broad pedigree. Usually no more than 2 sons per sire were retained in select lines and 1 son per sire was retained in control lines. In addition, in the early stages of introducing industry sires, genetic group values were poorly estimated, especially for maternal effects. Some progeny of introduced sires were initially retained and used, ignoring poorly estimated EBV.
Target EBV for maternal genetic weaning weight and direct genetic yearling weight were determined differently for purebred and composite populations. Target EBV for composites were set to be similar to mean EBV at the end of the GPU experiment. Many industry purebred populations had substantially increased yearling weight and milk EPD from 1978 to 1992. Purebreds in the GPU experiment were not selected for growth during this period (Gregory et al., 1991a). Therefore, target EBV for purebred populations were set to increase genetic levels of growth and maternal weaning weights in the purebred populations toward those in industry. Growth and maternal weaning weight targets in purebreds were increased each year. The intention was to increase EBV for these traits by at least two-thirds of the change that occurred in industry herds from 1978 to 1992. Target EBV for birth weight was set to be in the same proportion to yearling weight EBV targets as the phenotypic proportion. Proportionality of control line birth weight and yearling weight EBV targets was applied only to the EBV means of selected animals and not to individual animal EBV. Maintaining the ratio of birth weight to yearling weight was done in an attempt to minimize changes in the genetic level of heifer calving difficulty in the control lines.
Heifer calving difficulty score maternal and direct EBV were weighted differently for selection in different populations. A "terminal" weighting of 0:1 maternal to direct was used in Charolais and MARC I. A "general purpose" weighting of 1:1 maternal to direct was used in Gelbvieh and MARC II. A "maternal" weighting of 2:1 maternal to direct was used in Angus, Hereford, and MARC III. Weightings were applied to unstandardized direct and maternal calving difficulty score EBV.
Purebred industry bulls used in select and control lines were further selected for above-average EPD for yearling weight and milk to make the experimental populations closer to industry growth and milk EPD. Industry Angus, Charolais, and Hereford bulls used in select lines had low birth weight EPD. Gelbvieh bulls were chosen for desirable calving ease EPD. Industry purebred bulls used in control lines were selected for birth weight EPD that was proportionally similar to yearling weight EPD.
Figure 1 illustrates the main features of selection among MARC II males born in 1997 after weaning and after yearling EBV calculations. A common target EBV for yearling weight was used for select and control lines. Control and select line bulls selected as yearlings had similar average yearling weight EBV. Control line EBV for birth weights of selected yearling bulls were near their target EBV.
|
Selection of heifers born in 1996 and 1997 was relaxed. Heifers retained for breeding were randomly selected within sire. Relaxation of selection allowed the evaluation of 2-yr-old heifer calving difficulty in 1998 and 1999 unbiased by possible phenotypic effects of selection for EBV.
Statistical Analyses of EBV
Differences between select and control line EBV were evaluated for animals born in 1998 and 1999. The following model was fitted by using the MIXED procedure (SAS Inst. Inc., Cary, NC):
 |
where Yklm is the direct EBV for calving difficulty score, birth weight, weaning weight, or postweaning gain, or maternal EBV for calving difficulty score, birth weight, or weaning weight; LINEk is the select or control; POPl is the population; and 
klm is the residual error for the mth animal in the kth line and lth population. Population and the interaction of line x population were considered random. Line differences (select line minus control line) were tested for significance with the F-test. The Satterthwaite method was used to determine degrees of freedom.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
). An intentionally high cow replacement rate was desirable and resulted in 31% of calves born to 2-yr-old dams. More heifers were scored for calving difficulty, making more selection among females and a decreased generation interval possible. This was expected to result in more accurate heifer calving difficulty EBV, increased genetic change, and more accurate measurement of phenotypic trends.
|
. Select line sires were used an average of 2.1 yr and had 19.7 progeny. Control line sires were used for 1.2 yr and averaged 8.7 progeny. Fewer progeny per sire in the control line partially compensated for expected larger genetic drift from having fewer cows in the control lines. In purebred populations, 27% of select line sires and 25% of control line sires were introduced from industry. Introduced sires were 15% of all sires used in Gelbvieh, 18% in Charolais, 34% in Angus, and 38% in Hereford. Differences in use partially reflected differences among industry purebred population EPD trends from 1977 through 1992 and therefore expected differences between experimental and industry populations. More industry sires were used in Angus and Herefords, because the differences between industry and the experimental populations were thought to be the largest in those breeds.
|
Select and Control Differences
Average EBV trends show (Figure 2) that most selection objectives were achieved. Average EBV for the final 2 yr (1998 and 1999) are shown in Table 3. The average difference between lines (select minus control) in calving difficulty score direct EBV was –1.06. The direct EBV difference for the highly correlated trait, birth weight, was –3.5 kg. Line differences for weaning weight direct and maternal EBV and yearling weight direct EBV were small and not significant.
|
|
,b), and that there are fewer maternal expressions of heifer calving difficulty score than expressions of birth weight, which is highly correlated with direct heifer calving difficulty. For these reasons, variation in maternal EBV was much less than direct EBV. Therefore, even a 2:1 maternal-to-direct EBV weighting did not result in much change in maternal EBV.
Another important reason for no average difference in maternal EBV for calving difficulty score was the introduction of industry sires into the purebred populations. Sires introduced for the 1992 breeding season did not have any daughters calving until 1995. The industry genetic groups typically had too few 2-yr-old daughters calving to accurately estimate their group effects until the middle or later years of the experiment. Averages by population and line for maternal and direct EBV for 1998 and 1999 (Table 4) showed that composites with nonzero maternal EBV weightings had maternal EBV differences in the expected direction, but purebred maternal populations were not consistent. Introducing industry sires resulted in inaccuracies in maternal EBV in the first half of this relatively short selection experiment and affected selections based on these EBV.
|
) showed large changes in maternal weaning weight and direct yearling weight EBV for purebred populations, especially Angus and Hereford, but small changes in composites. Industry Angus and Hereford showed large changes in growth and milk EPD from 1977 to 1992. Changes in 2 x EPD for milk and yearling weight during that period were 7 and 40 kg for Angus (http://www.angus.org/sireeval/genetic.html, accessed 3/7/2006) and 5 and 32 kg for Hereford (http://www.hereford.org/Acrobat/Perf/S06_Trend.pdf, accessed 3/7/2006), respectively. Changes in the experimental populations (Table 5) were 80% or more of the 1977 to 1992 changes in industry Angus and Hereford for these traits.
|
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
; Rutledge et al., 1973; Atchley et al., 1997). These antagonistic selection experiments generally used phenotypic index selection, a single stage of selection, and had discrete generations.
In cattle, most reported experiments have used single-trait selection or index selection with a mildly antagonistic objective. Exceptions to this include 2 selection experiments for birth and yearling weights (MacNeil et al., 1998; MacNeil, 2003). In the first experiment, a 2-stage independent culling process was used to first select for below-average birth weight and then for high yearling weight. This resulted in small changes in birth weight and lower birth and yearling weights than did single-trait selection for yearling weight. The second experiment based selection on the index of Dickerson et al. (1974; yearling weight – 3.2 x birth weight) compared with an unselected control line. This selection resulted in small positive increases in birth weight and moderate positive increases in yearling weight. Arnold et al. (1990) reported the results of selecting 5 low-birth-weight sires and 4 high-birth-weight sires, all with similar high yearling weights, based on industry EPD. Progeny had differences in birth weight and yearling weight similar to their EPD. The results reported in Table 4
and Figure 2 clearly show antagonistic average responses of decreased birth weight and increased yearling weight.
This experiment is a practical application of strongly antagonistic selection in beef cattle when using currently available breeding value estimation technology and resources. It incorporates the multiple sources, generations, and stages of selection using breeding values estimated from multiple, correlated traits. This experimental strategy is not appropriate for estimating realized heritabilities or genetic correlations in the way that model organisms and some livestock have been used. Instead, the lines are best used to verify the usefulness of EBV, to estimate correlated responses in traits with low heritabilities, and to evaluate potentially undesirable, nonlinear responses correlated to large genetic changes in primary traits.
Sorensen et al. (2003) recommends using control lines when estimating response to selection, even if likelihood analyses are to be used. Both select and control lines were included in this experiment, but control lines were not random. Instead, control lines were selected toward specific EBV goals to make desired contrasts with select lines. Control lines had many fewer cows than did select lines. However, AI and semen from GPU and industry still allowed many sires to be used. Variability of EBV means was larger in the control line (Figure 2), but trends were still clear.
Estimated breeding value differences resulting from selection clearly showed that cattle can be simultaneously selected for increased growth and less calving difficulty. The resulting select and control lines differed by approximately 0.8 genetic SD for calving difficulty score and 0.9 genetic standard deviations for birth weight. On average, yearling weight increased by more than 0.5 genetic standard deviation, but exceeded 1 genetic standard deviation in Angus and Herefords. Differences between lines should be large enough to evaluate correlated changes in the 4 main traits and other traits and to identify potential limitations of genetically decreasing calving difficulty score and birth weight. Unlike many examples of genetic differences in calving difficulty and birth weight, differences between these lines should not be partially or completely confounded with subsequent growth rates.
The primary approaches for improving genetic potential for beef cattle production are selection among populations, mating systems that utilize complementarity and heterosis, and selection within populations. All these tools should be used to enhance production. Different production traits are more easily addressed with different approaches. The antagonistic relationship between postnatal growth and birth weight among breeds limits selection among breeds as an effective tool for increasing growth and reducing calving difficulty, except for the use of Bos indicus dams, which can reduce the birth weight of Bos taurus-sired calves (e.g., Amen et al., 2007). Heterosis increases growth, birth weight, and calving difficulty (Gregory et al., 1991a,b) and is not an effective means for dissociating calving difficulty, birth weight, and growth. Use of complementarity by mating terminal sire breeds to older, maternal breed cows and by mating specialized sire lines (or breeds) selected for direct calving ease to heifers and young cows does reduce the conflict between faster growth and larger calves with more difficult births. Even so, the relationship between birth weight and postnatal growth among breeds limits the choice of terminal sire and maternal breeds that can be used. Within-breed selection for increased growth and reduced calving difficulty is clearly possible and is an approach that should be very useful for interbreeding and rotational mating systems.
The most important result from this study is the clear demonstration of substantial genetic variation in heifer calving difficulty score and birth weight that is partially independent of yearling weight and that can be selected by using multiple-trait EBV technology. An ad hoc EBV selection procedure was used in this study, but any selection procedure (e.g., index, linear programming, culling levels, etc.) based on EBV should produce similar changes if it results in EBV selection differentials that are proportional to the changes in this experiment. However, weighting of EBV for use by the cattle industry should be based on economic principles. This study demonstrates that appropriately weighted EBV should produce the desired and predicted results even though traits might be complicated or antagonistic. This result is expected to apply generally to other antagonistically correlated traits important to beef cattle production, such as carcass quality and carcass yield.
|
Conclusions |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
|
Footnotes |
|---|
2 Mention of trade name, proprietary product, or specified equipment does not constitute a guarantee or warranty by the USDA and does not imply approval to the exclusion of other products that may be suitable.
3 Corresponding author: gary.bennett{at}ars.usda.gov
Received for publication November 11, 2007. Accepted for publication April 14, 2008.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONConclusionsLITERATURE CITED |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. L. Bennett, R. M. Thallman, W. M. Snelling, and L. A. Kuehn Experimental selection for calving ease and postnatal growth in seven cattle populations. II. Phenotypic differences J Anim Sci, September 1, 2008; 86(9): 2103 - 2114. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) represent males selected to remain intact after weaning and after yearling EBV calculations. Open squares (
) represent intact males not selected.