| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Animal Breeding and Genetics, Swedish University of Agricultural Sciences, PO Box 7023, 750 07 Uppsala, Sweden
2 Swedish Dairy Association, Hållsta, PO Box 1146, 631 80 Eskilstuna, Sweden
Corresponding author: E. Strandberg; e-mail: Erling.Strandberg{at}hgen.slu.se.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: genetic correlation • heritability • health • dairy cattle
Abbreviation key: LSCS = lactation average somatic cell score, MAST = clinical mastitis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Selection has traditionally focused on production traits. Today it is generally accepted that undesirable genetic relationships exist between production and health disorders, including mastitis (e.g., Rauw et al., 1998). According to several studies, milk production is unfavorably genetically correlated with both clinical mastitis and SCC (e.g., Emanuelson et al., 1988; Nielsen et al., 1997; Rupp and Boichard, 1999; Heringstad et al., 2000; Castillo-Juarez et al., 2002; Hansen et al., 2002), although some authors have reported favorable genetic associations between production and SCC in later parities (Pösö and Mäntysaari, 1996; Haile-Mariam et al., 2001a).
The heritability of clinical mastitis is low, especially when analyzed with linear models (Pösö and Mäntysaari, 1996; Rupp and Boichard, 1999; Lassen et al., 2003). Owing to the higher heritability of SCC and its high genetic correlation with clinical mastitis, it can be used for indirect selection to improve mastitis resistance (Mrode and Swanson, 1996; Heringstad et al., 2000). However, selection has been proven to be most efficient when information on clinical cases and SCC are combined (e.g., Philipsson et al., 1995). In the Swedish national genetic evaluation, bulls receive breeding values for clinical mastitis, based on information on clinical mastitis (veterinary treatments and culling due to mastitis) and SCC in first lactation daughters (Svensk Mjölk, 1999).
Mastitis is, however, not only a problem in first lactation. Actually, both mastitis frequency (Pösö and Mäntysaari, 1996; Nielsen et al., 1997) and level of SCC (Da et al., 1992; Reents et al., 1995; Nielsen et al., 1997) increase with increasing parity. Ideally, genetic evaluation for mastitis resistance also would include information from later lactations. Depending on the genetic parameters, multiple lactation records can be considered either as different traits in a multi-trait model or as repeated manifestations of the same trait in a single-trait repeatability model (Da et al., 1992; Reents et al., 1995).
The primary objectives were to estimate heritabilities of, and genetic correlations between, clinical mastitis and lactation average somatic cell score (LSCS) and to estimate genetic correlations between these udder health traits and production traits in the first 3 lactations of Swedish Holstein cows. A further aim was to estimate genetic correlations for udder health traits across lactations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
. To make bivariate analysis computationally feasible, the total data set (1) was split into 2 smaller data sets of equal size (2 and 3) by assigning herds randomly to either data set. All known pedigree information of the cows was traced back as far as possible, resulting in relationship matrices of 539,919, 288,809, and 286,171 animals for data sets 1, 2, and 3, respectively.
|
Definition of Traits
Mastitis, SCC, and production were defined in the same way as in the Swedish national genetic evaluation. A case of mastitis (MAST) was defined as a veterinary-treated clinical mastitis (with or without teat injury) from 10 d before to 150 d after calving, or culling for mastitis within that period. The restricted time period was used to reduce bias due to culling. Mastitis was defined as a binary trait distinguishing between cows with at least one reported case during the defined period (1) and cows without cases (0). Lactation average somatic cell score (LSCS) was the arithmetic mean of monthly test day SCC from 5 to 150 d after calving, expressed in 10,000 cells/mL, and transformed to a logarithmic scale with base 10 before averaging. Production of milk, fat, and protein (kg) was based on completed 305-d lactations. For interrupted lactations of >45 d length, and ongoing lactations of >100 d length, production was extended to 305-d yield. Real 305-d yield was analyzed for lactations of >305 d length, and for completed lactations of <305 d length total production in that lactation was analyzed without extension (i.e., MILK, FAT, PROT).
Statistical Analysis
(Co)variance components were estimated by REML, and analyses were performed with the DMU package (version 6, release 4) developed by Madsen and Jensen (2000). Both convergence criteria were set to 10–6. Estimates of heritabilities were derived from univariate analyses on data set 1 and estimates of correlations between traits and between lactations for the same trait were averages from 2 bivariate analyses on data sets 2 and 3. The following linear animal model was used for the production traits:
 |
where yijkl is the observation; hyi is the fixed effect of ith herd by year of calving; ymj is the fixed effect of jth year by month at calving; agek is the fixed effect of kth age in months at calving (one month per class); al is the random effect of lth animal; b1 is the fixed regression coefficient on the proportion of heterosis of animal l (Hetl); b2 is the fixed regression coefficient on the proportion of North American Holstein of animal l (Holl); b3 is the fixed regression coefficient on days open of animal l (DOPl); and eijkl is the random residual effect. The same model, without the regression on days open, was used for the udder health traits. Random effects were assumed to have zero means and the covariance structure for bivariate analysis was:
 |
where A is the additive relationship matrix, I is the identity matrix, and the indices represent the 2 traits in the bivariate analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
. Figure 1 shows the cumulative relative frequencies of mastitis from 10 d before to 150 d after calving in the first 3 lactations. Mastitis frequency was highest at calving and in the beginning of all lactations, and it was higher for first-lactation cows than for cows in later lactations. About 46% of all cases up to 150 d in first lactation occurred within 10 d after calving, whereas the same level was reached after 44 and 34 d in the second and third lactations, respectively. For all lactations, the total number of cases within 150 d of lactation constituted about 60 to 65% of all cases in completed lactations. Thus, the opportunity period (–10, 150) captures a large part of all cases. Even though only a few percent of all cases occurred before calving, it has been previously shown that heritability increases when these days are included in the registration period for mastitis (Heringstad et al., 1997).
|
). No clear pattern could be seen for the effect of year and month at calving (January 1995 to December 2000) on mastitis frequency, whereas LSCS tended to be lower and production traits higher for cows calving during the last 6 mo of a year (results not shown).
|
. The effect of heterosis was less MAST, lower LSCS, and higher production. The effect of North American Holstein was more cases of MAST, higher LSCS, and higher production.
|
Heritabilities and Correlations
Heritabilities and correlations between traits within parities are provided in Table 4. Heritabilities of MAST were low: 0.03 for first lactation and 0.01 for later lactations. For LSCS, heritabilities were considerably higher than those of MAST, but here also estimates decreased slightly with increasing parity, from 0.14 to 0.10. Heritabilities of milk, fat, and protein production were of moderate size, ranging from 0.23 to 0.36.
|
Genetic correlations between milk production traits in the first 3 parities varied between 0.11 and 0.87, although most estimates were moderate to high. The highest correlation for all lactations was between milk and protein, whereas the correlation between milk and fat was lowest. The strength of the correlations declined with increasing parity, especially between milk and fat. Estimated environmental correlations were all high, about 0.9 to 1.0, and again the highest estimates were between milk and protein.
Estimated genetic correlations between milk production traits and udder health traits were all positive, which indicates an undesirable relationship. The magnitude of the correlations, however, varied considerably between traits and lactations. For instance, MAST was most strongly correlated to milk (0.26 to 0.45) and least to fat. The correlation between MAST on one hand and milk or protein on the other was highest in the second lactation, whereas the correlation between MAST and fat decreased with increasing parity down to near zero in the third lactation. However, the estimates of correlations are averages of 2 bivariate analyses, and for the third lactation, the genetic correlations between MAST and all 3 production traits for the 2 subsets (2 and 3) differed considerably, with estimates for data set 2 being close to zero or slightly negative. The same was true for the correlation between MAST and fat in the second lactation. The corresponding standard errors for these genetic correlations were high. Environmental correlations between MAST and production traits were low and negative for all traits and lactations, ranging between –0.06 and –0.13.
Estimated genetic correlations between LSCS and production traits were lower than the corresponding correlations between MAST and production traits. The highest estimates of LSCS were found for the first lactation (0.17 to 0.23). In later lactations the correlations in our study decreased to about 0.1 and 0.2 for LSCS and milk or protein, respectively, and they were close to zero for LSCS and fat. In similarity with the estimates between MAST and production in the third lactation, the correlations between LSCS and production in the third lactation are less precise and are averages based on 2 estimates that differed markedly. Environmental correlations between LSCS and production traits were estimated at about –0.2 in first lactation and somewhat weaker in later lactations.
The correlations across lactations for the same trait are given in Table 5. Estimated genetic correlations of MAST across the first 3 lactations were all above 0.7. The highest estimate was between second and third lactation (>0.9), whereas the lowest was between first and third lactation. However, the average value between first and third parity is calculated from 2 estimated genetic correlations (0.46 and 0.98) that differed considerably. One of the 2 bivariate analyses for the genetic correlation between parities 2 and 3 had convergence problems, even though parameter estimates changed very slowly, and it was therefore interrupted after 50 iterations. Environmental correlations between MAST across lactations were positive but close to zero.
|
For production traits, estimated genetic correlations across lactations were high, around 0.9 between first and later lactations, and near unity between the second and third lactations. Environmental correlations across lactations for all production traits were about 0.3 to 0.4.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Estimated heritabilities of LSCS (0.10 to 0.14) are in agreement with previously reported estimates. In a review, Mrode and Swanson (1996) reported estimates between 0.05 and 0.47, with weighted average heritabilities of SCC of 0.11 (SD 0.04) and 0.11 (SD 0.07) for first and later lactations, respectively. A later review reports estimates ranging from 0.08 to 0.19 (Heringstad et al., 2000), and more recent estimates are of similar size varying between 0.09 and 0.18 (Haile-Mariam et al., 2001b; Castillo-Juarez et al., 2002; Søndergaard et al., 2002; Mrode and Swanson, 2003; Ødegård et al., 2003). In most studies including later lactations, only a slight variation of heritability of SCC in various lactations was found. However, Da et al. (1992), for example, observed increases in heritability with increasing parity for the first 3 lactations (0.05 to 0.11), whereas Banos and Shook (1990) reported that heritability decreased with increasing parity for the first 3 lactations (0.14 to 0.11).
In our study, the heritability of all traits decreased with increasing lactation number. This was mainly an effect of increasing residual variances but also, in some cases, due to decreasing genetic variances. The lower heritability in later lactations could also partly be explained by selection in first parity. This was confirmed in the analyses with 2 lactations for the same trait, where estimated heritabilities were more similar for different lactations than from the univariate analysis, especially for production traits (for example, for protein in lactation 1 to 3, average heritabilities estimated from bivariate analysis were 0.29, 0.27, and 0.27, whereas heritabilities from univariate analysis were 0.31, 0.25, and 0.23).
Correlations Between Udder Health Traits
The high estimates of genetic correlations between MAST and LSCS (around 0.7 to 0.8) found in this study are in the upper range of reported estimates cited in the literature (Mrode and Swanson, 1996; Rupp and Boichard, 1999; Heringstad et al., 2000), although estimates close to unity have been found (Lund et al., 1994). Heringstad et al. (2000) reviewed 7 studies based on Nordic field data, where estimates of genetic correlations between clinical mastitis and SCC ranged from 0.3 to 0.8, with an average of 0.6. Not many studies have taken later lactations into account. In our study, genetic correlation was higher in third lactation. This is in agreement with the results from Pösö and Mäntysaari (1996), where the largest increase was between first (0.4) and later parities (0.6 and 0.7 in the second and third lactations, respectively) and Nielsen et al. (1997), who found the highest correlation for the third lactation for one of their data sets and no clear difference in the other data set.
Correlations Between Udder Health Traits and Production Traits
In the literature, the genetic correlations between clinical mastitis and production traits have generally been unfavorable. This corresponds to results in our study, where estimates ranged from 0.01 to 0.45, with higher estimates found for first and second parities. Estimates of genetic correlation between mastitis susceptibility and milk yield based on Nordic data ranged from 0.24 to 0.55 (Heringstad et al., 2000). Other reported estimates for first-lactation cows, between clinical mastitis on one hand and milk, protein, or fat yield on the other, also ranged from 0.2 to 0.5 (Emanuelson et al., 1988; Uribe et al., 1995; Nielsen et al., 1997; Van Dorp et al., 1998; Heringstad et al., 1999; Rupp and Boichard, 1999; Hansen et al., 2002). In agreement with our results, Rupp and Boichard (1999) found the lowest correlation between clinical mastitis and fat (0.15) and the highest between clinical mastitis and milk (0.45) in first-lactation Holstein cows.
Estimated genetic correlations between LSCS and production traits for the first 3 lactations ranged from near zero to 0.2, with the strongest correlations for first lactation. These are in the upper range of previously reported estimates. Mrode and Swanson (1996) found, for first lactation, a weighted average genetic correlation between SCC and milk, fat, and protein yields of 0.14 (SD 0.04 to 0.05). More recent estimates for first lactation were between close to zero and 0.3 (Pösö and Mäntysaari, 1996; Charfeddine et al., 1997; Luttinen and Juga, 1997; Nielsen et al., 1997; Pösö et al., 1997; Rupp and Boichard, 1999; Castillo-Juarez et al., 2002). In similarity with previous studies, we found a lower genetic correlation between LSCS and fat yield than between LSCS and milk or protein yield (Charfeddine et al., 1997; Rupp and Boichard, 1999; Castillo-Juarez et al., 2002).
The strength of genetic correlations between LSCS and production traits in our study decreased with increasing parity, although estimates remained positive. Other authors reported that the genetic correlation between SCC and milk production, changed from positive, thus unfavorable, in the first lactation, to negative in later lactations (Banos and Shook, 1990; Pösö and Mäntysaari, 1996; Haile-Mariam et al., 2001a). Two possible explanations for the changes in genetic correlation between parities have been given (Banos and Shook, 1990). First, partly different genes may affect SCC in first vs. later lactations because different pathogens may be mainly responsible for the mastitis cases. Second, it has been argued that culling practices, especially during first lactation, that remove low-producing cows with high occurrence of mastitis and high levels of SCC may have an influence on genetic correlations. However, we would not expect that culling practice to give the observed change in genetic correlation, rather the opposite.
Correlations Across Lactations for Udder Health Traits
For both MAST and LSCS the highest genetic correlations across lactations were between second and third parities and the lowest between first and third parities, probably due to the longer time interval between them. For MAST (0.7 to 0.9) this was in agreement with results from Pösö and Mäntysaari (1996) and Nielsen et al. (1997). Although the size of our estimates were very similar to those estimated by Pösö and Mäntysaari (1996), they were lower overall compared with the estimates by Nielsen et al. (1997) (0.9 to 1.0). The high genetic correlation between parities 2 and 3 could be used as an argument for a multi-trait model with first and later lactations as separate traits.
Estimates of LSCS across lactations (0.8 to near unity) were similar to previously reported estimates. Mrode and Swanson (1996) summarized genetic correlations across lactations for SCC and found simple averages of 0.77, 0.76, and 0.87 between lactations 1 and 2, lactations 1 and 3, and lactations 2 and 3, respectively. More recent studies report genetic correlations around 0.7 to 0.9 between first and second parities, 0.7 to 0.8 between first and third parities, and 0.9 to near unity between second and third parities (Pösö and Mäntysaari, 1996; Boichard and Rupp, 1997; Nielsen et al., 1997; Boettcher et al., 1998; Mrode and Swanson, 2003). Our results suggest that LSCS should be considered as the same trait genetically for lactations 2 and 3, and as a separate but highly correlated trait for lactation one. Thus, based on genetic correlations only, a multi-trait model with first and later lactations as separate traits can be proposed.
Accuracy in Selection for Mastitis Resistance
To compare the accuracy in selection for mastitis resistance when selection is based on MAST, LSCS, or a combination of both measures, selection index theory was used. Accuracy is defined as the correlation between the true breeding goal, which in this case is freedom from clinical cases of mastitis, and the indices, being composed of LSCS, MAST, or LSCS + MAST. Parameters assumed were those estimated in first lactation. The progeny group sizes used were 50, 100, and 150 daughters. The resulting accuracies are shown in Table 6.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Mastitis frequency and level of LSCS increased with increasing parity. Therefore it is important that selection programs seek to improve mastitis resistance in all parities. Waiting for information from later lactations before selecting young bulls would create a prolonged generation interval, which is not desirable, and because genetic correlations between parities were relatively high (>0.7) for both MAST and LSCS, resistance in later lactations will be improved even if only first-lactation records are used. However, even with these rather high correlations, inclusion of later-parity information in the genetic evaluation would be expected to enhance accuracy somewhat, through inclusion in pedigree information. Given the estimated correlations, a multi-trait model with first and later lactations as separate traits can be suggested for both MAST and LSCS.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Received for publication October 1, 2003. Accepted for publication March 3, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Ouweltjes, J. J. Windig, G. de Jong, T. J. G. M. Lam, J. ten Napel, and Y. de Haas The Use of Data from Sampling for Bacteriology for Genetic Selection Against Clinical Mastitis J Dairy Sci, December 1, 2008; 91(12): 4860 - 4870. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Heringstad, E. Sehested, and T. Steine Short Communication: Correlated Selection Responses in Somatic Cell Count from Selection Against Clinical Mastitis J Dairy Sci, November 1, 2008; 91(11): 4437 - 4439. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Holtsmark, B. Heringstad, P. Madsen, and J. Odegard Genetic Relationship Between Culling, Milk Production, Fertility, and Health Traits in Norwegian Red Cows J Dairy Sci, October 1, 2008; 91(10): 4006 - 4012. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Samore, A. F. Groen, P. J. Boettcher, J. Jamrozik, F. Canavesi, and A. Bagnato Genetic Correlation Patterns Between Somatic Cell Score and Protein Yield in the Italian Holstein-Friesian Population J Dairy Sci, October 1, 2008; 91(10): 4013 - 4021. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Negussie, I. Stranden, and E. A. Mantysaari Genetic Association of Clinical Mastitis with Test-Day Somatic Cell Score and Milk Yield During First Lactation of Finnish Ayrshire Cows J Dairy Sci, March 1, 2008; 91(3): 1189 - 1197. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Lund, G. Sahana, L. Andersson-Eklund, N. Hastings, A. Fernandez, N. Schulman, B. Thomsen, S. Viitala, J. L. Williams, A. Sabry, et al. Joint Analysis of Quantitative Trait Loci for Clinical Mastitis and Somatic Cell Score on Five Chromosomes in Three Nordic Dairy Cattle Breeds J Dairy Sci, November 1, 2007; 90(11): 5282 - 5290. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Ouweltjes, B. Beerda, J. J. Windig, M. P. L. Calus, and R. F. Veerkamp Effects of Management and Genetics on Udder Health and Milk Composition in Dairy Cows J Dairy Sci, January 1, 2007; 90(1): 229 - 238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Mark and P. G. Sullivan Multiple-Trait Multiple-Country Genetic Evaluations for Udder Health Traits J Dairy Sci, December 1, 2006; 89(12): 4874 - 4885. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. de los Campos, D. Gianola, and B. Heringstad A structural equation model for describing relationships between somatic cell score and milk yield in first-lactation dairy cows. J Dairy Sci, November 1, 2006; 89(11): 4445 - 4455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Carlen, U. Emanuelson, and E. Strandberg Genetic evaluation of mastitis in dairy cattle using linear models, threshold models, and survival analysis: a simulation study. J Dairy Sci, October 1, 2006; 89(10): 4049 - 4057. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Banos, M. P. Coffey, E. Wall, and S. Brotherstone Genetic relationship between first-lactation body energy and later-life udder health in dairy cattle. J Dairy Sci, June 1, 2006; 89(6): 2222 - 2232. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Heringstad, D. Gianola, Y. M. Chang, J. Odegard, and G. Klemetsdal Genetic associations between clinical mastitis and somatic cell score in early first-lactation cows. J Dairy Sci, June 1, 2006; 89(6): 2236 - 2244. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Gonda, Y. M. Chang, G. E. Shook, M. T. Collins, and B. W. Kirkpatrick Genetic Variation of Mycobacterium avium ssp. paratuberculosis Infection in US Holsteins J Dairy Sci, May 1, 2006; 89(5): 1804 - 1812. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. D. Norman, J. R. Wright, R. L. Powell, and P. M. VanRaden Impact of Maturity Rate of Daughters on Genetic Ranking of Holstein Bulls J Dairy Sci, September 1, 2005; 88(9): 3337 - 3345. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Powell, A. H. Sanders, and H. D. Norman Accuracy and Stability of National and International Somatic Cell Score Evaluations J Dairy Sci, July 1, 2005; 88(7): 2624 - 2631. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Heringstad, Y. M. Chang, D. Gianola, and G. Klemetsdal Genetic Association Between Susceptibility to Clinical Mastitis and Protein Yield in Norwegian Dairy Cattle J Dairy Sci, April 1, 2005; 88(4): 1509 - 1514. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Carlen, M. d. P. Schneider, and E. Strandberg Comparison Between Linear Models and Survival Analysis for Genetic Evaluation of Clinical Mastitis in Dairy Cattle J Dairy Sci, February 1, 2005; 88(2): 797 - 803. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
), two (
), and three (x) of Swedish Holstein cows.
a) on last line for each lactation. The subscripts are the approximated standard errors for the estimates.2