| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,
,4
* Department of Animal Science, University of Nebraska, Lincoln 68583-0908; and USDA-ARS, Roman L. Hruska U.S. Meat Animal Research Center,
and
Clay Center, NE 68933 and
and
Lincoln, NE 68583-0908
4 Correspondence:
A218 Animal Sciences (phone: 402-472-6010; fax: 402-472-6362; E-mail:
dvan-vleck1{at}unl.edu).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
Key Words: Genetic Correlation • Genetic Parameters • Growth • Heritability
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
. Hereford-Angus reciprocal crosses were produced in each cycle of the program. Some Angus and Hereford sires used in Cycle I were repeated as reference sires in following cycles to provide genetic ties across all four cycles.
|
Calves were born in the spring (March to early May) each year. Males were castrated within 24 h after birth. Calves were allowed to suckle and graze with their dam, and were creep fed from mid-July or early August until weaning in early October at about 200 d of age. The F1 females (Phase 2) in each cycle were retained to evaluate growth, age at puberty, reproduction, and maternal performance through mature ages. Heifers were managed to have their first calf at 2 yr of age. Details of postweaning management were presented by Laster et al. (1976, 1979) and Gregory et al. (1979) for Cycles I to III, and Cundiff et al. (1998) for Cycle IV. The first three cycles are summarized by Cundiff et al. (1986), Cycle IV is described by Cundiff et al. (1998) and Thallman et al. (1999), and all four are reported by Cundiff et al. (1993).
Cows were on improved pastures (April to November) of cool or summer mixtures of grass. During winter months (December to March), cows were maintained on grass and legume hay in Cycles I to III, and on corn silage and alfalfa hay in Cycle IV. The length of the breeding season was about 75 d with pregnancy diagnosed by palpation 60 to 75 d after the end of the breeding season. Cows were culled only if they failed to conceive in two successive years or for serious unsoundness (i.e., lameness, unsound udders, prolapse, cancer eye, emaciation). Cows were not culled on the basis of growth criteria. Yearling heifers were weighed at start and end of the mating season and when palpated for pregnancy in the fall. Thereafter, cows were weighed, measured for hip height, and scored for body condition four times each year. One measurement was taken during each season: 1) mid-May (spring), at the start of the breeding season; 2) the beginning of August (summer), at the end of the breeding season; 3) the end of October (fall), at palpation for pregnancy following weaning; and 4) the beginning of February (winter), prior to calving. Body condition score was on a subjective classification scale of nine points (Spitzer, 1986) from (1) extremely thin (emaciated) to (9) extremely fat (obese).
The number of records for each trait by age of cow and season of measurement are in Table 2. A total of 61,798 measurements of weight were available. As weight and condition score may be affected by physiological status of the cow, especially if the cow is rearing a calf or is pregnant, each record of a cow was assigned to one of four physiological codes corresponding to a combination of lactation status (1 = not lactating, 2 = lactating) and pregnancy status (1 = not pregnant, 2 = pregnant). The present study included only records of 24 mo or older cows with assigned physiological code. Original data included cows that were separated from the experiment to be sold or used in other studies. Some edits were needed to provide consistency of data within and across cycles. For Cycle I, records of cows up to 7 yr of age were used. For Cycle II, cows up to 8 yr of age were included. For Cycle III, cows with records up to 6 yr of age were considered. For Cycle IV, 6 yr of age was the cut off point because limited data were available for older cows. Three birth years (1986, 1987, and 1988) had data until 6 yr of age, whereas cows born in 1989 and 1990 had records until 5 and 4 yr of age respectively. For Cycle-II crossbred cows from the diallel experiment with Red Poll and Brown Swiss dams were excluded. Heterosis and transmitted effects for weight, height, and condition score from that experiment were reported for 2- to 7-yr-old cows (Dearborn et al., 1987). Gregory and Maurer (1991) reported results for those traits at 30, 42, and 54 mo of age.
|
The equation for the linear model was:
 |
where yis the vector of weights, heights, or condition scores; ß is the vector of fixed effects; and a andc are vectors of random cow additive genetic and permanent environmental effects with incidence matrices X, Za, and Zc, respectively, and e is the vector of residual effects. From the variance components, three parameters were defined: 1) heritability, 
; 2) fraction of permanent environmental variance, 
; and 3) fraction of temporary environmental variance, 
. In addition repeatability (r) was defined as the sum of h2 and c2.
For bivariate analyses, the model was:
 |
where 1 and 2 represent different seasons or cow ages. Local convergence was declared when the variance of -2 times the log likelihood in the simplex was less than 10-8. Restarts were done to increase the chance of finding a global maximum rather than a local maximum, with convergence declared when -2 times the log likelihood did not change to the third decimal place. When this rule was not met after 10 restarts for multivariate analyses, convergence was declared if the (co)variance estimates did not change after the second decimal point. Standard errors were computed with the "delta" method based on the average information matrix at convergence (e.g., Dodenhoff et al., 1998).
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
for all the traits. Heritability estimates were high for cow weight (0.49), cow weight adjusted for condition score (0.54), and for cow height (0.68), but relatively low for cow body condition score (0.16). Corresponding estimates for the fractions of permanent environmental variance were 0.16, 0.13, 0.07, and 0.14, respectively. Repeatability estimates were 0.65, 0.66, 0.75, and 0.30, respectively. These results suggest that selection should be effective for mature weight and height and that repeated records for cow weight and height may have little value. Condition score, however, had low heritability and moderate repeatability. With weight adjusted for body condition score, estimates of variance components did not change much. The phenotypic variance decreased by approximately 9%, the estimate of heritability increased by about the same magnitude, and the fraction of permanent environmental variability decreased approximately twice as much. Repeatability was similar (0.65 and 0.67) for both traits. Most of the difference in estimates of heritability between weight and weight adjusted for condition score seems to be explained by differences in variance of permanent environmental effects. Repeated records of weight may represent different environmental and physiological conditions affecting cow fatness that were not completely adjusted by fixed effects in the model but were better removed when weights were adjusted for condition score. Alternatively, since most of relationships were through sires, differences in heritability between the two models may represent direct genetic differences of sires for fatness and/or milking ability as expressed by differences in condition scores of their daughters. Similar results have been found in other studies (Meyer, 1995; Johnston et al., 1996).
|
The overall estimate of heritability for cow weight in this study agrees with the extensive review of Koots et al. (1994), who reported a pooled estimate of heritability of 0.50 for mature weight from 25 studies. More recent estimates using ANOVA methods include: 0.52 from the average of seven bivariate analyses, using a modified Henderson method with a sire-maternal grand sire model (Bullock et al., 1993). Estimates from REML of heritability for cow weight ranged from 0.49 to 0.68 according to a study involving Angus-Hereford reciprocal crosses in New Zealand (Morris et al., 1987). Northcutt and Wilson (1993) reported estimates of 0.48 from univariate analyses and of 0.45, 0.51, and 0.51 from bivariate analyses with birth, weaning, and 365-d weight, respectively, from 5- to 12-yr-old Angus cows. Meyer (1995) reported heritability estimates for cow weight in Australian Herefords and a composite breed (Wokalups) ranging from 0.30 to 0.59 depending on the breed and age.
Heritability estimates for cow weight adjusted for body condition score have ranged from 0.38 to 0.76 using ANOVA based methods (e.g., Morrow and Marlowe, 1966; Choy et al., 1996, 1998b) and from 0.45 to 0.79 with REML (Northcutt and Wilson, 1993; Johnston et al., 1996; Wilson, 2000). Predicted asymptotic mature weight and mature weight represent the same genetic trait as shown by Meyer (1995), who found a genetic correlation of 0.99 between them in Australian Hereford and Wokalups.
Estimates of repeatability for cow weight have been between 0.45 and 0.88 using least squares and the paternal half-sib (PHS) method (e.g., Brinks et al., 1962; Fitzhugh et al., 1967; Williams et al., 1979; Field et al., 1988) and between 0.57 and 0.85 using REML (Morris et al., 1987; Meyer, 1995; Johnston et al., 1996).
Cow height was more heritable (h2 = 0.68) and repeatable (r = 0.75) than cow weight (Table 3
), which agrees with the few previous reports. Estimates of heritability for cow height using ANOVA methods have ranged from 0.54 to 0.75 in Angus, Hereford, and crossbred cows (Brown and Franks, 1964; Neville et al., 1978; Choy et al., 1996, 1998a, b). Estimates of repeatability have ranged from 0.78 to 0.82 (Choy et al., 1996, 1998a,b). Estimates by REML of heritability from field records (Angus) were 0.83 and 0.85 from bivariate analyses with weight and weight adjusted for condition score (Northcutt and Wilson, 1993; Wilson, 2000). Meyer (1995) reported estimates of heritability (0.19 to 0.50) that were less than those reported here; however, the corresponding repeatability estimates (0.73 to 0.83) were similar to the estimate from this study (0.75).
Cow body condition score had the lowest heritability (0.16) and repeatability estimates (0.30) of the traits in the present study (Table 3). Previous estimates using ANOVA methods have ranged from 0.13 to 0.51 (e.g., Brinks et al., 1964; Marlowe and Morrow, 1985; Choy et al., 1996, 1998a, b) for heritability, whereas repeatability estimates have varied from 0.34 to 0.85 (Benyshek and Marlowe, 1973; Williams et al., 1979; Choy et al., 1996, 1998a, b). Meyer (1995), using REML, reported estimates of heritability ranging from 0.12 to 0.16, and of repeatability from 0.20 to 0.25, depending on breed (Hereford or Wokalups) and ages included in the analyses. Johnston et al. (1996) obtained estimates of heritabilities of 0.21, 0.14, and 0.17 and repeatabilities of 0.32, 0.52, and 0.41 for Angus, Hereford and Polled Hereford cows (n = 7,536) included in the Australian BREEDPLAN.
Estimates of variance components by season of measurement are presented in Table 4. Means and phenotypic variances changed among seasons for the two traits related with cow weight, as well as for condition score, which was expected because seasons represent not only different climate and environmental conditions (especially nutritional level), but also different physiological status of the cows, particularly regarding pregnancy and lactation, which may affect fat deposition. Cows were heavier and had better condition scores in summer and fall, as expected. Height did not change across seasons, as anticipated for a skeletal trait in adult animals. Heritability estimates followed the overall age results (Table 3) and were similar across seasons. Estimates of heritabilities were slightly higher for cow weight than for cow weight adjusted for condition score. Heritability estimates for spring (at calving) and fall (at weaning) weights have been reported to be similar using the PHS method (Brinks et al., 1962, 1964; Fitzhugh et al., 1967). Morris et al. (1987) compared REML estimates of heritability of weight taken four times during the year of Angus-Hereford reciprocal crosses in New Zealand. Heritability estimates (0.49 to 0.60) and their differences among seasons were similar to the estimates found in the present study.
|
. Estimates of phenotypic variances for weight increased with age, suggesting that age accounted for a systematic increase in size. Coefficients of variation were much lower for height (
3%) than for weight (9%). Phenotypic variances did not change much, and tended to decrease at later ages for height in agreement with previous research. Brown et al. (1956a,b) indicated that skeletal growth was nearly completed at 3 yr of age and that further increases in weight were due primarily to increases in muscle and fat deposition. Phenotypic variances for condition score were similar over the range of 2 to 8 yr of age. With cow weight adjusted for condition score, phenotypic variance relative to cow weight unadjusted for condition score decreased with increase in age, by 11.5% at 2 yr to 26.5% at 8 yr. Heritability estimates tended to increase from 2 to 5 yr of age, then decreased until 7 yr, and increased again at 8 yr of age for weight, adjusted weight, and height. However, the estimates did not show a significant trend across ages, ranging from 0.47 to 0.58, 0.51 to 0.63, and 0.59 to 0.72, respectively. Heritability estimates for condition score were about 0.20 for all ages, except for an estimate of 0.51 for 8-yr-old cows, which may be due to the limited number of cows.
|
In general, estimates of heritability and repeatability from the present study agree with previous reports. Cow weight and height had intermediate to large estimates, whereas estimates for body condition score were low to intermediate.
Bivariate Analyses
Results for bivariate analyses are summarized in Tables 6 to 10. Many correlations were close to the boundary (1.00), slowing the search procedure and convergence. Estimates of heritability for weight (Table 6), weight adjusted for body condition score (Table 7), and height (Table 8) were remarkably similar to those from univariate analyses. Genetic correlations between ages were always large for weight (0.92 to 1.00) and height (0.98 to 1.00), tending toward unity between consecutive ages, and decreasing slightly for ages more separated for both traits. Correlations between permanent environmental effects were close to one for successive ages, but decreased with increasing age differences to 0.56 and 0.55 (weights) and 0.72 (height) between extreme ages (2 to 6 yr). Phenotypic correlations followed the same trend as genetic and permanent environmental correlations, ranging from 0.63 to 0.73 for weight, 0.62 to 0.72 for weight adjusted for condition score, and 0.72 to 0.83 for height. Phenotypic correlations were moderate to high (0.56 to 0.86) among seasons of measurement for weights.
|
|
|
|
) were slightly greater than those from univariate analyses. Heritability estimates agreed well with those from univariate analyses for age classes, but were more erratic (0.11 to 0.20) for season classes. Genetic correlations were high, except between extreme ages (0.67). Permanent environmental correlations ranged from 0.31 to 0.88. All genetic and permanent environmental correlations between seasons converged to 1.00. Phenotypic correlations were intermediate among ages (0.23 to 0.39) and among seasons (0.28 to 0.62).
|
Brinks et al. (1964) reported a large estimate of genetic correlation (0.93) between weights taken in spring and fall for range Hereford cows. Jeffery and Berg (1972) also reported a high genetic correlation (0.87) between cow weights at two times of the year (postcalving and weaning of the calf) in Canadian purebred and crossbred animals.
Results from bivariate analyses between traits over all data are summarized in Table 10. Phenotypic variances agreed closely with values from univariate analyses for all traits with the exception of cow weight, for which the variance was 10% higher with the bivariate analyses. Heritability estimates were the same as for the univariate analyses for all traits. Estimates of genetic correlations were high between size and weight measurements (0.80 and 0.86 between weight and height and between weight adjusted for condition score and height). Estimates of genetic correlations were moderate (0.43) between weight and condition score and near zero (-0.04) between height and condition score. Permanent environmental correlations were intermediate between height and weight (0.55 and 0.61 for weight and weight adjusted for condition, respectively), high between weight and condition score (0.79), and relatively low between height and condition score (0.26). In general, results reported here indicate no antagonistic relationships between the traits. Selection would be effective for either weight or height, and would produce important correlated responses for both measurements of growth. Selection for condition score would not be as effective, but would exhibit a correlated response with weight. Selection for height would not affect condition score under the nutritional resources of this study.
Previous studies of Cycles I to III of the GPE program have shown estimates similar to the present study. Jenkins et al. (1991) reported high genetic correlations within breed (0.73), between breeds (0.87), and total (0.76) between asymptotic mature weight (estimated from records adjusted for condition score) and mature height (7 yr). Estimates of 0.73, 0.59, and 0.70 and 0.59 and 0.58 were found between 500-d weight and mature height. Data from nine purebred and three generations (F1, F2, and F3) of composite populations from the Germplasm Utilization program at MARC yielded estimates of genetic correlations of 0.74, 0.48, and 0.00 between weight and height, weight and condition score, and height and condition score of 2-yr-old cows, and 0.47, 0.56, and -0.14 at 5 yr of age (Gregory et al., 1995).
Least squares estimates of phenotypic correlations between cow weight and height have been reported to be from 0.06 to 0.89 (Brown et al., 1956a,b,c; Nelsen et al., 1985; Thompson et al., 1983; Naazie et al., 1989; Choy et al., 1996, 1998a, b), but most estimates were intermediate to high, averaging 0.59, the same estimate obtained in the present study. Estimates of genetic correlations have been relatively high (0.68 to 0.79), as reported by Choy et al. (1996, 1998a,b). Estimates of correlations by REML between cow weight and height also agree with the present study. Northcutt et al. (1992) and Northcutt and Wilson (1993) reported estimates of 0.58, 0.56, and 0.78 for phenotypic, environmental, and genetic correlations, respectively, for Angus field data. Meyer (1995) obtained estimates for the corresponding correlations of 0.59 and 0.55, 0.88 and 1.00, and 0.76 and 0.67 in Australian Herefords and Wokalups. Correlations between cow weight adjusted for condition score and height were reported to be 0.54 to 0.61 and 0.66 to 0.85 for phenotypic and genetic correlations (Choy et al., 1998b; Northcutt and Wilson, 1993; Wilson, 2000).
Estimates of genetic correlations between weight and conditions have ranged from 0.20 to 0.76, averaging 0.46 (Brinks et al., 1964; Marlowe and Morrow, 1985; Choy et al., 1996, 1998a, b). Meyer (1995), using REML, obtained estimates for phenotypic, environmental, and genetic correlations between weight and condition score of 0.60 and 0.58, 0.85 and 0.74, and 0.78 and 0.64 for Australian Hereford and Wokalups, respectively.
Estimates of the genetic correlation between cow height and cow condition have been 0.02 to 0.30, averaging 0.11 (Choy et al., 1996, 1998a,b) for Angus cows. Estimates of correlations by REML between height and condition score include an estimate of 0.10 for the environmental correlation for Angus (Northcutt et al., 1992). Meyer (1995) reported estimates of 0.19 and 0.04, -0.28 and 1.00, and 0.76 and -0.39 for phenotypic, environmental, and genetic correlations in Australian Hereford and Wokalups, respectively.
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
|
Footnotes |
|---|
2 Published as paper no. 13556, Journal Ser., Nebraska Agric. Res. Div., Univ. of Nebraska, Lincoln 68583-0908.
3 Current address: Facultad de Ciencias Veterinarias, Universidad Central de Venezuela, Apartado. 4563, Maracay 2101, Aragua, Venezuela.
Received for publication November 1, 2001. Accepted for publication July 25, 2002.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Roughsedge, P. R. Amer, R. Thompson, and G. Simm Genetic parameters for a maternal breeding goal in beef production J Anim Sci, October 1, 2005; 83(10): 2319 - 2329. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Arango, L. V. Cundiff, and L. D. Van Vleck Covariance functions and random regression models for cow weight in beef cattle J Anim Sci, January 1, 2004; 82(1): 54 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Arango, L. V. Cundiff, and L. D. Van Vleck Comparisons of Angus, Charolais, Galloway, Hereford, Longhorn, Nellore, Piedmontese, Salers, and Shorthorn breeds for weight, weight adjusted for condition score, height, and condition score of cows J Anim Sci, January 1, 2004; 82(1): 74 - 84. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Arango, L. V. Cundiff, and L. D. Van Vleck Breed comparisons of Angus, Charolais, Hereford, Jersey, Limousin, Simmental, and South Devon for weight, weight adjusted for body condition score, height, and body condition score of cows J Anim Sci, December 1, 2002; 80(12): 3123 - 3132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Arango, L. V. Cundiff, and L. D. Van Vleck Comparisons of Angus-, Braunvieh-, Chianina-, Hereford-, Gelbvieh-, Maine Anjou-, and Red Poll-sired cows for weight, weight adjusted for body condition score, height, and body condition score J Anim Sci, December 1, 2002; 80(12): 3133 - 3141. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Arango, L. V. Cundiff, and L. D. Van Vleck Breed comparisons of Angus, Brahman, Hereford, Pinzgauer, Sahiwal, and Tarentaise for weight, weight adjusted for condition score, height, and body condition score J Anim Sci, December 1, 2002; 80(12): 3142 - 3149. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
p), heritabilities (h2), and genetic (
a), permanent environmental (