| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Agriculture and Agri-Food Canada, Guelph, ON, Canada N1G 4T2
2 Canadian Dairy Network, Guelph, ON, Canada N1G 4T2
Corresponding author: A. Sewalem; e-mail: sewalem{at}cdn.ca.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: functional survival • Holstein • type trait
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Type traits have been used as indirect selection criteria for herd life (Short and Lawlor, 1992; Dekkers et al., 1994; VanRaden and Wiggans, 1995; Weigel et al., 1998; Cruickshank et al., 2002). Type traits are recorded relatively early in life, most often in the first lactation, and are more heritable than longevity (Yazdi and Schaeffer, 2001; Kadarmideen and Wegmann, 2003), which makes selection relatively more efficient. To get reliable and direct information for sires regarding the longevity of their daughters, it is necessary to wait until a minimum number of daughters are culled or die. Moreover, these evaluations may be available too late to be useful in breeding programs. Hence, genetic evaluations for direct longevity information based on number of culled cows should be combined with indirect information based on early predictors such as type traits. Knowledge of genetic relationships between type and longevity are required, and, therefore, a proper identification of type traits to be used as early predictors is essential. In Canada, the breeding goal has emphasized high production combined with superior conformation to support such production levels, and considerable improvement on type traits has been made compared with other countries. Because type classification data are already recorded, their use as an indirect predictor for longevity would also be very cost effective.
Survival analysis using a Weibull proportional hazards model can offer a better fit to survival data because of its ability to account for censored records properly. This may increase precision by accounting for differences in days of productive life among cows that survive for the same number of lactations. Survival analysis also accounts for the skewed distribution of survival data. Time-dependent variables can be used for the survival analysis to model the effects of environmental factors accurately.
The objective of the present study was to evaluate the impact of linear descriptive type traits on functional longevity of Canadian Holstein cows.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
, and further information can be found at the Holstein Canada Web site (http://www.holstein.ca).
|
 |
where 
(t) is the hazard of a cow, i.e., her probability of being culled at time t given that she is alive just before t; 0,s(t) = lambda;
(t)–1 is the Weibull baseline hazard function with scale parameter and shape parameter; and t is the time in days from one calving to the next calving for each stratum; ß contains the possibly time-dependent covariates affecting the hazard, with xm'(t) being the corresponding design vectors, and u is a vector of random variables with associated incidence vector zm'.
The fixed covariates included in the model were as follows: time-dependent effect of stage of lactation in days (1 = 0 to 80; 2 = 81 to 235; 3 = >235); effect of year and season of calving (year of calvings from 1985 to 2003 and season of calvings were January to March, April to June, July to September, and October to December); effect of season of production with the same definition as seasons of calving; effect of the annual change in herd size with 3 classes (decreasing = a decrease in herd size of <–5%, nearly unchanged = no appreciable change 
–5 to 
10%, and increasing = an increase in herd size of >10%); effect of the type of milk recording supervision with 3 classes (unsupervised, supervised, and unknown, i.e., records that do not fulfill the minimum criteria set by the milk recording agency); effect of age at first calving in months; and effects of milk, fat, and protein yields. The latter effects were calculated as within herd-year-parity deviations with 3 classes for each: low = cows producing >0.4 SD below the herd-year-parity average, average = cows producing between 0.4 below and 0.6 SD above the herd-year-parity average, and high = cows producing above 0.6 SD of the herd-year-parity average. Each type trait was evaluated separately by including it as a covariate on the model.
The random effects included were the effect of herd-year-season class, which was assumed to follow a log gamma distribution, and the genetic effect of the cow’s sire, which was assumed to follow a multivariate normal distribution with mean zero and variance A
2s, where 2s is the variance among sires and A is the relationship matrix. A sire variance of 0.046 (Sewalem et al., 2003) was used in the analysis.
The following composite type traits were analyzed: final score, mammary system, feet and legs, dairy character, rump, frame/capacity, fore udder, and rear udder. The linear type traits analyzed are presented on Table 1
. Analyses were performed for each type trait one at a time; hence, 31 analyses were performed on the same data according to the model previously described.
The analyses, using a Weibull proportional hazards model, were performed using the Survival Kit Version 3.12 (Ducrocq and Solkner, 1998). One baseline hazard function, 0,s(t), was defined for each lactation (subscript 0 designates a baseline hazard, and subscript s relates to stratum). Detailed description of the model and survival analysis of longevity data in dairy cattle on a lactation basis were provided by Ducrocq (2002), Roxstrom et al. (2003), and Sewalem et al. (2003). The overall influence of each type trait on functional survival was assessed using the likelihood ratio test.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
shows the relative contribution of each type trait (composite and linear traits) to the likelihood (–2logL); this was determined by comparing the full model (with one particular type trait) to the reduced model (without any of the type traits). Final score was by far the most important single trait with respect to longevity, followed by mammary system, rear and fore udders, and feet and legs (Figure 1). Among the linear type traits, udder traits had the strongest effect on the survival of cows. Fore udder attachment, udder texture, udder depth, rear udder attachment height, rear udder attachment width, and median suspensory were the udder traits having the strongest relationship with longevity of cows. Among the feet and legs traits, bone quality and heel depth had the largest influence on functional survival. Stature and size had the least influence on the survival of cows. Only traits related to udder and feet and legs contributed at least 20% of the relative contribution of final score to the likelihood. The average Weibull parameters () were 1.6, 1.5, 1.4, 1.4, 1.4, and 2.5 for lactations 1, 2, 3, 4, 5, and 6+, respectively. The log 
value ranged from 8 to 20.
|
Composite Traits
Figure 2 shows a clear linear relationship between final score and longevity. Cows with an extremely low final score of <65 (i.e., classified as poor) had a risk of being culled that was 3.7 times more than that of cows scoring 80 points; whereas cows with final scores of >89 (i.e., classified as excellent) had a risk of being culled that was 0.28 times that of 80-point cows. Final score was the most important type trait. Similar results using the same methodology were also reported by Vollema and Groen (1998) for Dutch dairy cows, Schneider et al. (2003) for Quebec Holstein cows, and Caraviello et al. (2004) for US Holstein cows. Boettcher et al. (1997) used a linear regression approach and reported that overall conformation had the largest effect on survival through first lactation of Canadian Holsteins. Caraviello et al. (2003) reported, however, that final score was not the most important type trait that influences the functional longevity of US Jersey cows, perhaps because the Jersey breed is likely to have much less voluntary culling of cows with low final scores.
|
shows the relationship between risk of relative culling and mammary system, fore udder, and rear udder. Mammary system was the second most important trait next to final score for functional longevity in this analysis. Low scores (classified as poor) were 2 times more likely to be culled compared with the reference score of 10, which relates to a cow classified as good plus 1. On the other hand, cows with scores >15 (i.e., classified as excellent) had a risk of being culled that was 0.5 times less than cows with the reference score of 10. Figure 3 also shows that cows with low scores for rear and fore udder were >3 times more likely to be culled compared with cows with intermediate scores. The higher relative culling risk for category excellent 3 (i.e., score of 18) compared with excellent 2 (i.e., score of 17) as shown in Figure 3 is due to few numbers of cows in the former category. Figure 4 shows the relationship between relative culling risk and feet and legs, rump, dairy character, and capacity. Relative culling risks for feet and legs also showed a clear relationship with a cow’s survival. Low scores were associated with less risk compared with the higher scores. Capacity and rump angle also showed the same trends as feet and legs; however, the magnitude of risk of culling for these traits was much lower. Figure 4 also shows the influence of dairy character on functional survival. The relative culling risk for cows with low scores was>2.6 times that of cows with an intermediate score for this trait. A positive strong relationship of mammary system and feet and legs with longevity was also reported by Boettcher et al. (1997), Larroque and Ducrocq (2001), and Schneider et al. (2003).
|
|
. Fore teat length, fore teat placement, rear teat placement, and udder depth exhibit an intermediate optimum as they relate to longevity. On the other hand, fore udder attachment, median suspensory ligament, rear udder attachment height, rear udder attachment width, and udder texture show a clear linear relationship with functional longevity. Cows with high scores for these traits had higher chances of surviving than cows with low scores. Udder traits had an important influence on the culling decisions. This might have been due to their influence on the reduced susceptibility to mastitis and other infectious diseases. Cows with nearly center fore teat placement are more likely to survive culling than cows with extremely inside or extremely outside fore teat placement, which has also been observed in other studies (Caraviello et al., 2003; Schneider et al., 2003). Cows with extremely close rear teats were more likely to be culled compared with cows with extremely wide rear teats, which was also found by Buenger et al. (2001), Larroque and Ducrocq (2001), and Schneider et al. (2003). However, Caraviello et al. (2003) reported an intermediate optimum for rear udder height and read udder width. This result might have been due to breed differences or trait definitions. Low scores for median suspensory ligament or depth of cleft for rear udder and fore udder were associated with higher risk of culling than the other scores. Cows with extremely fleshy udders were more likely to be culled compared with cows with soft, pliable, elastic udders that collapse well after milking. Cows with extremely deep or extremely shallow udder depth from hock to the floor of the udder were also associated with higher risk of being culled.
|
. Cows that are extremely short, small, and narrow-chested with a shallow body depth had a higher risk of being culled compared with cows in an intermediate class and, hence, a shorter length of productive life. As for size, Table 3 also shows that cows with a score of 9 were essentially equivalent to a score of 2, indicating that extremely large cows have a relative risk of culling similar to that of extremely small cows. A moderate positive relationship between longevity and body type traits was reported by Boettcher et al. (1997). Size and stature did not have a strong relationship with functional survival (Buenger et al., 2001; Caraviello et al., 2003). On the other hand, Schneider et al. (2003) found that taller and bigger cows had better chances of surviving than cows in other classes. However, Mahoney et al. (1986) reported that larger cows are more predisposed to displaced abomasums than small cows. Rogers et al. (1999) did not find a significant difference, indicating that genetically larger body size or more strength would be advantageous from a disease perspective.
|
. A clear linear relationship between angularity and longevity was observed, indicating that the rounded, nonangular cows were 2.47 times more likely to be culled compared with the intermediate score of 5; extremely sharp cows had a 1.28 times better chance of surviving than cows with intermediate scores. However, Buenger et al. (2001) and Caraviello et al. (2004) did not find a linear relationship between angularity and longevity. Moreover, Buenger et al. (2001), after correcting for milk production, found no strong relationship between angularity and longevity of cows. The varying relationships between angularity and longevity in different countries may be partly due to differences in the trait definition across countries. In Canada, the appraisal of angularity is based on the degree of angular shape at the chine and the fore rib, also accounting for the spring of rib. The cow receiving high marks for angularity has angularity with substantial, wide well-sprung, and open ribs. The frail, flat-sided cow is coded down for angularity, as she lacks this spring of rib.
Rump traits.
The relationships between risk of relative culling and linear score for loin strength, pin width, and rump angle are shown in Table 4. Cows with weak strength of vertebrae between back and rump showed a higher culling risk compared with cows with strong loins. Cows with low or high scores for rump angle (representing extremely high or low pin bones relative to the height of hook bones, respectively) were more likely to be culled compared with the optimum linear score of 5. No strong relationship was observed between pin width and relative culling risk except that cows with extremely narrow pins were 1.18 times more likely to be culled compared with cows that were scored 5 (Table 4). Rump traits are associated with calving difficulty, and cows with intermediate rump angle are preferred. Similar findings were reported by Buenger et al. (2001) and Schneider et al. (2003). Moreover, Rogers et al. (1999) found a favorable moderate positive genetic correlation of 0.39 between rump angle and foot and leg diseases. However, Caraviello et al. (2003, 2004) did not find strong relationships between rump traits and longevity in US Holstein and Jersey cows.
|
shows the relationship between the risk of relative culling and linear score for feet and legs traits. There was a slight linear relationship for bone quality and rear legs rear view with relative culling risk. Cows with low scores were more likely to be culled compared with those having high scores for these traits. The relationship between foot angle and rear legs side view with longevity showed an intermediate optimum, indicating that cows with low and high scores were more likely to be culled compared with those having an intermediate score of 5. The importance of feet and legs traits to the functional survival of cows found in the present study corroborates the findings of Boettcher et al. (1997), Buenger et al. (2001), and Caraviello et al. (2004). Extremely coarse bone, extremely shallow heel depth, extremely low foot angle, and extremely straight or curved rear legs led to a decreased functional survival of cows. Results from this study are from a model that includes a sire genetic effect. Buenger et al. (2001) and Larroque and Ducrocq (2001) argued that inclusion of the genetic effect in the model would bias the effect of type trait’s phenotype effect on culling risks. In this study, a phenotypic analysis of each trait (excluding the sire effect in the model) was also analyzed, and results were compared with the present study (not shown). Neither the estimates of breeding values nor significance tests changed. Schneider et al. (2003) also found no differences between phenotypic and genetic analysis of type traits and its impact on the functional survival of cows.
Previous studies by Dekkers et al. (1994), Boettcher et al. (1997) using different methodology, and Schneider et al. (2003) using the same approach as the present study but data only from Quebec Holstein cows reported the impact of type traits on survival of cows for Canadian Holsteins. The completeness and volume of data and the advantage of methodology in the present study reflect the relationship between type traits and functional herd life more precisely than previous studies. The present study also rendered an estimate of the contribution of each type trait to the hazard function, or instantaneous risk of relative culling, in Canadian Holstein cows. Both the previous and present studies assume that culling for type traits is involuntary. Breeders of Canadian Holsteins have always placed more emphasis on type traits than Holstein breeders in other countries. Therefore, voluntary culling for type traits was much more likely for cows in this data set than for cows in other studies of Holstein populations globally. The extent to which culling for type is voluntary, and the association between type traits and fertility and/or disease incidences such as reproductive problems, metabolic disorders, and foot and leg disease, would be of interest in the future.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Received for publication February 6, 2004. Accepted for publication July 24, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Sewalem, F. Miglior, G. J. Kistemaker, P. Sullivan, and B. J. Van Doormaal Relationship Between Reproduction Traits and Functional Longevity in Canadian Dairy Cattle J Dairy Sci, April 1, 2008; 91(4): 1660 - 1668. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Miglior, A. Sewalem, J. Jamrozik, D. M. Lefebvre, and R. K. Moore Analysis of Milk Urea Nitrogen and Lactose and Their Effect on Longevity in Canadian Dairy Cattle J Dairy Sci, December 1, 2006; 89(12): 4886 - 4894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Wiggans, L. L. M. Thornton, R. R. Neitzel, and N. Gengler Genetic Parameters and Evaluation of Rear Legs (Rear View) for Brown Swiss and Guernseys J Dairy Sci, December 1, 2006; 89(12): 4895 - 4900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Sewalem, F. Miglior, G. J. Kistemaker, and B. J. Van Doormaal Analysis of the relationship between somatic cell score and functional longevity in canadian dairy cattle. J Dairy Sci, September 1, 2006; 89(9): 3609 - 3614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Wall, I. M. S. White, M. P. Coffey, and S. Brotherstone The Relationship Between Fertility, Rump Angle, and Selected Type Information in Holstein-Friesian Cows J Dairy Sci, April 1, 2005; 88(4): 1521 - 1528. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Sewalem, G. J. Kistemaker, and B. J. Van Doormaal Relationship Between Type Traits and Longevity in Canadian Jerseys and Ayrshires Using a Weibull Proportional Hazards Model J Dairy Sci, April 1, 2005; 88(4): 1552 - 1560. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), fore udder (
), and rear udder (
) (relative culling rate for category 10 was set to 1).
), and dairy character (