| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL GENETICS |
,1
* Associazione Nazionale Allevatori Bovini di Razza Piemontese, 12061 Carrù, Italy;
and
Animal Breeding and Genetics Group, Wageningen University, 6700 AH Wageningen, The Netherlands; and
and
Department of Animal Science, University of Padova, 35020 Legnaro, Italy
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
Results indicate that, due to the existence of antagonistic relationships between the investigated traits, specific selection strategies need to be studied.
Key Words: Beef Cattle • Calving Performance • Growth • Fleshiness • Genetic Correlations
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
). Biologically, this trait is affected by a direct component, related to calf size, and a maternal component, mainly related to the size of the pelvic area of the dam (Philipsson et al., 1979). Recently, a number of authors provided estimates of genetic parameters for both direct and maternal calving performance in beef cattle (Varona et al., 1999; Carnier et al., 2000), indicating that exploitable genetic variation exists for these traits. Few studies investigated the relationship between calving performance and other traits that are of importance for efficiency of beef production such as growth rate, muscularity, and slaughter performance (Renand, 1985a,b; Gregory et al., 1995). Moreover, in most cases, only the direct effect of calving performance was considered when estimating covariances with other traits.
The Piemontese is an important breed specialized for beef production in Italy. Piemontese cattle exhibit double muscling (Grobet et al., 1998) and relatively high levels of dystocia. The economic relevance of calving performance and possible unfavorable effects induced by selection for beef production traits suggest inclusion of this trait in the breeding goal of the Piemontese population. Current breeding goals include daily gain, live fleshiness, bone thinness, and direct and maternal calving performance (Carnier et al., 2000; Albera et al., 2001), but covariances between beef traits and calving performance are assumed to be null.
The aim of this study was to estimate genetic correlations between direct and maternal calving performance and beef traits for Piemontese cattle.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
A total of 1,602 records of young bulls enrolled in the performance-testing program from 1989 to 2002 at the central station of the Italian Piemontese Cattle Association were available.
At the station, bulls are weighed monthly from 2 up to 12 mo of age and daily gain (DG) is computed as the slope of the regression of live weight on age at weighing. For bulls tested after 1991 (1,292 bulls), scores for live fleshiness and bone thinness attributed by trained classifiers were also available. Fleshiness was appraised on six different body sites using a nine-score linear system: withers width (WW), shoulder muscularity (SM), limbs width (LW), limbs thickness (LT), thigh muscularity (rear view, TM), and thigh profile (side view, TP). Bone thinness of the shin bone (BT) was also evaluated with the same scale. Each animal was independently scored by three classifiers. Details on the performance testing procedure of Piemontese bulls can be found in Albera et al. (2001).
A recording system of calvings according to the level of difficulty was adopted in 1989. Five scores are used: 1 (unassisted delivery), 2 (assisted easy calving), 3 (difficult calving), 4 (Cesarean section), and 5 (embryotomy). Information about presentations of calf was not systematically recorded; therefore, it was impossible to exclude calving records with abnormal presentations. To estimate covariances between beef traits and calving performance, only informative calving records were considered: birth records of tested bulls, of their sire and dam, half sibs, progeny, and of the progeny of their half sibs. Furthermore, birth records of calves having a tested bull or its male half-sib as maternal grandsire were used. This editing had little effect on the data structure for herds using AI as all the bulls selected for AI were also tested for beef traits on station. Because the use of AI is widespread in the Piemontese breed, analyzed records were approximately 70% of the total records available. Discarded records were mainly from herds using natural service bulls, which were excluded from the analysis because of the lack of connectedness with young bulls tested for beef production traits.
Differences in the incidence of dystocia and in the magnitude of estimated genetic parameters suggested the need to treat calving performance in heifers and cows as different traits (Carnier et al., 2000).
Incomplete records; records with missing sire, dam, maternal grandsire or granddam; and records from twinning births or from very small herds (less then 30 records over 14 yr) were removed.
After editing calving records and beef traits records were merged to form two datasets: Dataset 1 included DG, live fleshiness traits, bone thinness of station tested bulls and calving records in the first parity; Data-set 2 included the same information for beef traits associated with calving records in later parities. In Dataset 2, cows included as dams were required to have at least two calving records as later parities.
Observations on DG, live fleshiness, and bone thinness of calves not tested at station were treated as missing values. Similarly, for station-tested bulls without their own birth record, observations on calving score were set to missing.
All available pedigree information (on average, 4.5 generations of ancestors) was used to set up the numerator relationship matrix among animals.
After editing procedures, Datasets 1 and 2 contained 30,763 and 80,474 records, respectively, and the corresponding pedigree files had 86,270 and 115,605 records. Characteristics of the datasets are reported in Table 1.
|
The model used for the analysis of Dataset 1 was as follows:
 |
where y is a vector of observations, b is a vector of fixed effects, ud is a vector of random additive genetic direct effects, um is a vector of random additive genetic maternal effects, e is a vector of random residual, and X, Zd, and Zm are known incidence matrices relating observations to b, ud, and um. Superscripts denote observations and model terms related to calving performance in the first parity (c) and the beef trait analyzed jointly (b). For random effects, assumed means were null and variances were as follows:
 |
 |
where 
2d is the additive direct genetic variance, 2m is the additive maternal genetic variance, d is the additive genetic covariance between direct effects, dm is the additive genetic covariance between direct and maternal effects, 2e is the random residual variance, A is the numerator relationship matrix, I is an identity matrix, and 
is the Kronecker product operator.
Because a very low number of bulls tested on station were born from heifers (see Table 1
), residual covariances between calving performance of heifers and beef traits were assumed to be zero.
Dataset 2 was analyzed with a model similar to that used for Dataset 1, but a permanent environmental random effect was included for calving performance:
 |
where y, b, ud, um, e, X, Zd, and Zm have the same meaning as in the former model, up is a vector of random permanent environmental effects and Zp is an incidence matrix relating observations on calving performance to their respective random permanent environmental effects. Superscripts indicate calving performance in later parities (c) and the beef trait (b). Random components of the model were assumed to have null means and variances as follows:
 |
where 2d, 2m, d, dm, 2e, A, and I are defined as in the former model, 2p is the random permanent environmental variance, and e is the random residual covariance. In this analysis, residual covariance has been considered not null because most of the bulls tested for beef traits also had their own birth record.
Heritabilities of calving performance were derived from estimated (co)variances as in Carnier et al. (2000). Correlations between additive direct or maternal additive genetic effects on calving performance and additive genetic effects on a beef trait have been computed as follows:
 |
Nongenetic Fixed Effects
For calving performance in the first parity (Dataset 1), the vector of fixed effects included herd, year-season of calving, sex of the calf, age of the dam at calving, and the interaction between sex of the calf and age of the dam. For calving performance of cows (Dataset 2) effects of herd-year-season, age of the dam within parity, and sex of the calf were considered. As the number of small herds was not trivial, the herd-year-season effect could not be considered in the model used for first parity; in the analysis of Dataset 2 it was fitted only for large herds (herds with more than 200 calving records in 14 yr). For medium-sized herds (from 100 to 200 calving records) or small herds (less then 100 calving records), a herd-year or a herd effect was included, respectively. Two seasons of calving were defined, from November to April and from May to October. The age of the dam at calving was classified as follows: eight classes (21 to 37 mo) for heifers and 50 classes within parity (seven parities) for cows.
For beef traits, fixed effects in the model were as in Albera et al. (2001). For DG and BT, they included the contemporary group of animals on test (154 levels) and the parity of the dam treated in classes (four levels). For live fleshiness traits, the linear regression effect of the weight of bulls at scoring on the fleshiness score was also fitted. As individual classifiers have changed over time, it was impossible to include the effect of the classifier in the model adopted for the live fleshiness and bone thinness scores.
Because of missing genotype information for most animals, the effect of the genotype at the myostatin locus (Grobet et al., 1998) was not considered in the model both for calving performance and beef traits. However, genotyping of the Piemontese sires selected for artificial insemination in the last 30 yr revealed that all of the bulls were homozygous for the mutated allele (A. Albera, unpublished data). Because the rate of AI is high in Piemontese cattle, the genotype at the myostatin locus cannot be considered as a source of variation for these traits. All analyses were performed using average information REML (Gilmour et al., 2002).
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
. Heifers experienced more calving problems than adult cows; the amount of Cesarean sections was over three times higher for first vs. for later parities, and also difficult calvings were two times more frequent in heifers than in cows. Despite the very low frequency of calving records scored as embryotomy, merging these records with Cesarean sections was not considered because of the difference in the relative economic values (Albera et al., 1999).
|
). The use of a threshold model has been shown to lead to possible bias in the estimates of variance components (Moreno et al., 1997) when the number of progeny per sire is limited and one or more of the fixed effects included in the model have a large number of levels with a small number of observations within level. Furthermore, some problems in the estimation of fixed effects might also arise as a consequence of the extreme category problem, due to the occurrence of all observations for a class of a fixed effect in the same category of the response variable (Misztal et al., 1989). In such a situation, the use of linear models has been recently suggested as an acceptable solution to overcome data structure problems in the genetic evaluation of calving performance (Phocas and Laloë, 2003), and this approach is currently adopted in the Piemontese breed (Carnier et al., 2000).
Heritabilities
Estimates of (co)variances, heritabilities, and genetic correlations between direct and maternal genetic effects for calving performance in heifers and cows are presented in Table 3. Estimates obtained in different bivariate analyses were consistent and, as a result, pooled estimates are presented. Variances and heritability estimates were similar to those obtained from Carnier et al. (2000) for Piemontese cattle.
|
. Genetic covariances between direct and maternal effects were negative; as a consequence, the corresponding genetic correlations were also negative, –0.38 and –0.33 in the first and later parities, respectively, but their magnitude was smaller in comparison with estimates of Carnier et al. (2000). Although a sample of all available calving records was used, results indicate that the sampled data represented the complete population well. Differences with Carnier et al. (2000) can be ascribed to the use of different models: in this study a multiple trait model has been used and effects due to selection on beef traits occurring in an earlier stage are accounted for. The effect of selection on the first calving performance on estimated genetic parameters for calving performance in subsequent parities could not be considered.
Estimated variances and heritabilities for beef traits are reported in Table 4. Estimated parameters for beef traits obtained in bivariate analyses with calving performance of heifers were consistent with those from analyses with calving performance of cows. In agreement with results by Albera et al. (2001), DG was the beef trait with the highest estimated heritability. Live fleshiness and bone thinness were moderately heritable traits with estimates ranging from 0.29 to 0.47. In comparison with the current study, Albera et al. (2001) reported higher heritabilities for WW, SM, and LW, and lower estimates for TM and TP. These differences might be related to the use of bivariate models that also consider calving performance. It should also be noted that the amount of information on beef traits available for this study was almost doubled. This was not expected to affect the estimates of genetic parameters but only the magnitude of standard errors. For live fleshiness, which was subjectively scored by classifiers who changed over time, estimates might also differ due to the inclusion of new data, even though preliminary analyses showed a limited heterogeneity of variance across classifiers. The same change could not be observed for DG, probably because of the objective nature of weight data.
|
and 6, respectively.
|
|
; Eriksson et al., 2002).
Daily gain showed a positive genetic relationship with direct calving performance. Genetic correlations ranged from 0.42 in heifers to 0.50 in cows, providing evidence that animals having faster growth are likely to generate progeny experiencing more problems at birth. Because birth weight is known as the most important factor affecting the birth ability of calves (Meijering, 1984), the genetic correlation between DG and direct calving performance indicates that genes controlling growth before birth are also partly involved in the expression of postnatal growth. Few studies dealt with the relationship between calving performance and traits related to beef production. In a literature survey, Koots et al. (1994) reported an average genetic correlation between postweaning growth and direct calving ease similar to that of this study. Bennett and Gregory (2001), in a study involving several parental beef breeds and composite populations, found a genetic correlation of 0.36 between postweaning gain and direct calving difficulty score across breeds. Lower correlations have been reported by Gregory et al. (1995).
Maternal calving performance was negatively associated with DG both in first and later parities; hence, calving ability is improved in females with higher growth rates. A possible biological explanation of this association might be related to the positive correlation between growth rate and mature weight of dams (Koots et al., 1994), which results also in a larger pelvic inlet of cows and in a better aptitude to calve. However, evidence of associations between calving ability of cows and their BW or size is not reported in the literature. On the basis of the magnitude of the estimated genetic correlations, selection for enhanced growth rate of animals is expected to affect direct calving performance more intensively than maternal calving performance, causing an increase in the incidence of dystocia as a correlated response. The same pattern has been found by Averdunk et al. (1987) and Bennett and Gregory (2001), who reported that genetic correlations of growth with calving ease were stronger for direct vs. maternal effect.
Genetic correlations between direct calving performance and live fleshiness traits were low to moderate. In the first parity, all correlations were positive, with the exception of that with WW, suggesting that as the muscularity of animals increases due to selection, births become slightly more difficult. Among fleshiness traits, those related to the development of animals in terms of width, LW, and TM, showed higher correlations with direct calving performance. In later parities, evidence of an association between live fleshiness and birth ability of calves was not clear because some of the correlations were slightly negative and in general their size was lower compared with the first parity. Because the dimension of the pelvic area of cows increases with age, it might be possible that calf conformation and shape do not act as a limiting factor in the calving process of adult cows. Similar results in beef cattle have been reported by Gregory et al. (1995), who found a genetic correlation of 0.33 between the muscle score of young bulls born from 2-yr-old dams and their calving difficulty score. When considering young bulls born from more mature cows (over 3 yr of age), the genetic correlation was lower and negative. Other literature reports indicated an inconsistent (Renand, 1985) or small positive association (Averdunk et al., 1987) between direct calving difficulties and traits expressing muscular development of young bulls.
Maternal calving performance showed a moderate positive association with all live fleshiness traits. Genetic correlations ranged from 0.06 to 0.33 in the first parity and from 0.18 to 0.30 in subsequent parities. Selection for increased muscularity is therefore likely to have a detrimental effect on the calving ability of females, probably due to a decrease in the pelvic inlet dimension. A phenotypic relationship between cows’ muscularity and pelvic area has not been demonstrated in normally muscled breeds (Meijering, 1984), but has been reported for double-muscled cattle (Hanset and Jandrian, 1979). The only study that could be found in the literature concerning genetic correlation between live muscularity of bulls and maternal calving performance of their daughters indicated a lack of relationship between these traits (Averdunk et al., 1987). Unlike DG, live fleshiness traits were more correlated with calving performance as a trait of the dam than as a trait of the calf.
Bone thinness was negatively correlated with direct calving performance. Animals with thick bones are born with relatively more calving problems. Genetic correlation was larger for later parities compared with the first parity (–0.38 vs. –0.17). Genetic correlations of maternal calving performance with bone thinness were positive both for heifers and cows. Because skeletal development is related to size and weight, animals with thin bones probably also tend to be lighter at birth and this could explain the favorable correlation with direct calving performance. For the same reason, calving ability of females is expected to be poorer in thin-boned cows because of their decreased body development. Furthermore, a positive association has been reported between bone thinness and live fleshiness in Piemontese (Albera et al., 2001), which might affect the dimension of dam’s pelvic area. Renand (1985), using French beef breeds, with data from two different progeny testing stations, reported inconsistent genetic correlations between skeletal development of young bulls and their birth difficulty ranging from positive to moderately negative depending on the station.
As expected, residual correlations of calving performance in second and later parities with all of the beef traits were close to zero because the traits were measured in different environments and only a few animals had observations in both traits.
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
1 Correspondence: Strada Trinità 32/a (phone: +39-0173-750791; fax: +39-0173-750915; e-mail: info{at}anaborapi.it).
Received for publication July 9, 2004. Accepted for publication August 10, 2004.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P. Gutierrez, F. Goyache, I. Fernandez, I. Alvarez, and L. J. Royo Genetic relationships among calving ease, calving interval, birth weight, and weaning weight in the Asturiana de los Valles beef cattle breed J Anim Sci, January 1, 2007; 85(1): 69 - 75. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
