Estimation of genetic parameters for linear and miscellaneous type traits in Holstein-Friesian dairy cattle

doi:10.1016/0301-6226(90)90078-K

Livestock Production Science

Volume 26, Issue 3, November 1990, Pages 177-192

https://doi.org/10.1016/0301-6226(90)90078-K Get rights and content

Abstract

Sixteen linear type traits (scored 1–9), total type score (10–99) and 33 miscellaneous characteristics (scored 0 or 1) were analysed on 73 940 first lactation Holstein-Friesian dairy cows, progeny of 748 sires. Heritabilities and correlations amongst the traits were estimated, and the heritabilities of the miscellaneous characteristics transformed from the binomial to the underlying normal distribution. In general, heritabilities for body and teat traits were higher than those obtained for leg, feet and udder traits. Phenotypic correlations between the 0/1 traits were mostly low, whereas the genetic correlations obtained were high between some traits, suggesting the feasibility of reducing the number of traits recorded. The incidence of a number of the miscellaneous characteristics was very low, and these traits could either be eliminated from the recording scheme or merged with other suitable traits. Incidences and heritabilities differed substantially among field officers (classifiers).

Résumé

On a analysé 16 caractères de type linéaire (notés de 1 à 9). Les notes globales de type (10–99) et 33 caractères divers (notés 0 ou 1) sur 73 940 vaches primipares Holstein-Friesian issues de 748 taureaux.

On a estimé l'héritabilité et les corrélations entre caractères et l'héritabilité des caractères divers a été transformée de la distribution binominale à la distribution normale sous-jacente. En général, l'héritabilité des caractères du coros et des travons a été supérieure à celle obtenue pour les caractères des membres, des pieds et de la mamelle. Les corrélations phenotypiques entre les caractères 0/1 ont été généralement faibles, tandis que les corrélations génétiques obtenues entre certains caractères étaient élevées. Ce qui suggere la possibilité de réduire le nombre des caractères enregistrés. La fréquence d'un certain nombre de caractères diversa été très faible: ils pourraient donc être éliminés des protocoles d'enregistrement ou fusionnés avec d'autres caractères appropriés. La fréquence et l'héritabilité des caractères ont différé notablement entre les notateurs.

Zusammenfassung

Sechzehn lineare Merkmale (Werte 1–9), Anzahl Werte total (10–99) und 33 gemischte Merkmale (Werte 0 oder 1) wurden an 73 940 erstlaktierenden Holstein-Friesian Milchkühen, Nachkommen von 748 Stieren, untersucht.

Geschätzt wurden die Heritabilitäten und Korrelationen zwischen diesen Merkmalen sowie die Heritabilitäten der gemischten Merkmale denen eine Normalverteilung zugrunde gelegt wurde. Allgemein waren die Heritabilitäten für die Körper- und Zitzenmerkmale höher als diejenigen für die Gliedmassen-, Klauen- und Eutermerkmale. Die phänotypischen Korrelationen zwischen den 0/1 Merkmalen waren meist tief. Die genetischen Korrelationen hingegen waren hoch zwischen einigen Merkmalen. Dies weist darauf hin, dass die Anzahl erhobener Merkmale reduziert werden könnte. Das Vorkommen gewisser gemischter Merkmale war sehr gering und diese Merkmale könnten entweder aus dem Erhebungsschema gestrichten oder mit anderen geeigneten Merkmalen zusammengefasst werden. Unter den Feldprüfern (Klassifizierern) fielen Vorkommen und Heritabilitäten sehr unterschiedlich aus.

References (18)

J.J. Colleau et al.
REML estimation of genetic parameters for type traits in Normande cattle breed
Livest. Prod. Sci.
(1989)
W.W. Foster et al.
Linear type analysis with genetic parameter estimation
J. Dairy Sci.
(1988)
D.A. Lawstuen et al.
Inheritance and relationships of linear type traits for age groups of Holsteins
J. Dairy Sci.
(1987)
G.W. Rogers et al.
Genetic correlations between survival and linear type traits measured in first lactation
J. Dairy Sci.
(1989)
L.D. Van Vleck
Estimation of heritability of threshold characters
J. Dairy Sci.
(1972)
V. Bowden
Type classification of dairy cattle: a review
Anim. Breed. Abstr.
(1982)
British Friesian Cattle Society
Introduction to linear type assessment
(1983)
J.-L. Foulley et al.
Empirical Bayes estimation of parameters for n polygenic binary traits
Genet., Sel., Evol.
(1987)
D. Gianola
On selection criteria and estimation of parameters when the variance is heterogeneous
Theor. Appl. Genet.
(1986)

There are more references available in the full text version of this article.

Cited by (33)

Direct and maternal genetic relationships between calving ease, gestation length, milk production, fertility, type, and lifespan of Holstein-Friesian primiparous cows
2013, Journal of Dairy Science
Citation Excerpt :
This was a consistent trend across a variety of dairy traits in the current study; thus, it cannot be ruled out that this phenomenon is an artifact of the fitted model. However, the fact that most of the genetic parameters yielded by this model were in line with estimates in literature (Brotherstone et al., 1990; Cue et al., 1990; Berry et al., 2004; Wall et al., 2005; López de Maturana et al., 2007a), which were obtained by a variety of statistical models, speaks strongly for its robust validity. It has also been demonstrated in the current study that not all previously observed associations of selection traits with CE and GL have a direct genetic basis.
As the emphasis in cattle breeding is shifting from traits that increase income toward traits that reduce costs, national breeding indices are expanding to include functional traits such as calving ease (CE). However, one issue is the lack of knowledge of genetic relationships between CE and other dairy traits. The same can be said about gestation length (GL), a potential novel selection trait with considerable heritabilities and possible genetic relationships with the calving process. This study aimed to estimate the genetic relationships between CE, GL, and other dairy traits of interest using a national data set of 31,053 primiparous cow performance records, as well as to separate direct and maternal genetic effects. Chosen dairy traits included fertility (calving interval, days to first service, nonreturn rate after 56 d, number of inseminations per conception), milk production (milk yield at d 110 in milk, accumulated 305-d milk yield, accumulated 305-d fat yield, accumulated 305-d protein yield), type (udder depth, chest width, rump width, rump angle, mammary composition, stature, body depth), and lifespan traits (functional days of productive life). To allow the separation of direct and maternal genetic effects, a random sire of the calf effect was included in the multi-trait linear trivariate sire models fitted using ASReml. Significant results showed that easily born individuals were genetically prone to high milk yield and reduced fertility in first lactation. Difficult calving primiparous cows were likely associated with being high-producing, wide and deep animals, with a reduced ability to subsequently conceive. Individuals that were born relatively early were associated with good genetic merit for milk production. Finally, individuals carrying their offspring longer were genetically associated with being wide and large animals that were themselves born relatively early. The study shows that it is feasible and valuable to separate direct and maternal effects when estimating genetic correlations between calving and other dairy traits. Furthermore, gestation length is best used as an indicator trait for lowly heritable calving traits, rather than as a novel selection trait. As estimated direct and maternal genetic correlations differ, we can conclude that genetic relationships between CE, GL, and traits of interest are present, but caution is required if these traits are implemented in national breeding indices.
Technical note: Prediction of liveweight from linear conformation traits in dairy cattle
2012, Journal of Dairy Science
Citation Excerpt :
Comparison criteria included the average absolute difference, root mean square error, and product-moment correlation. ( Co)variance estimates between liveweight and conformation traits were calculated with a multivariate model that included effects similar to the UK national genetic evaluation model for conformation traits (Brotherstone et al., 1990). Thus, year-season of inspection, age of cow at inspection (linear regression), stage of lactation at inspection (linear and quadratic regression), and month of calving were fitted as fixed effects.
The objectives of this study were to derive phenotypic and genetic prediction equations of liveweight from linear conformation traits, and estimate genetic and phenotypic parameters for these traits. Data pertained to 2,728 conformation and liveweight records of 613 cows in 1,529 lactations. Cows were raised at the Scottish Agricultural College research station and had calved between 2002 and 2010. Fifteen linear conformation traits were considered as predictors. To validate phenotypic predictions, the data set was randomly split into independent reference and validation subsets. Reference subsets were used to derive prediction equations with the use of a mixed model. Comparisons between predicted and actual liveweight in the validation subsets indicated that stature, chest width, body depth, and angularity could be used to derive phenotypic predictions of liveweight. Accuracy of these predictions was better for first-lactation than for all-lactation liveweight data. Significant genetic correlations between liveweight and the 4 predictor traits ranged from 0.49 to 0.76, and phenotypic correlations were 0.33 to 0.56. Estimated genetic (co)variances were used to develop prediction equations of animal genetic merit for liveweight from routinely calculated genetic evaluations for conformation traits.
Animal muscularity and size are genetically correlated with animal live-weight and price
2012, Livestock Science
Citation Excerpt :
Gutierrez and Goyache (2001) using Asturiana de los Valles beef cattle from Spain reported similar heritability estimates to those found in the present study for body condition score (0.33; Table 2) and all common skeletal type traits that were evaluated in this study (0.11 to 0.23; Table 2). Although heritability estimates of functional traits are lacking in beef cattle the lower heritability of the functional type traits (0.05 to 0.13; Table 2) in the present study compared to the skeletal and muscular type trait is consistent with previous studies in dairy cattle (Brotherstone et al., 1990; Meyer et al., 1987; Short and Lawlor, 1992). Heritability estimates for the linear type traits in the present study are also similar to other production traits in beef cattle such as live-weight (Meyer et al., 1993), auction price (Mc Hugh et al., 2011; Schierenbeck et al., 2008), and average daily gain (Crowley et al., 2010).
Animal price and live-weight contribute substantially to the overall profitability of beef and dairy production systems. However, animal data on price and live-weight are not routinely available on all animals since not all animals are sold or weighed during their lifetime. Previous studies have shown that linear type traits can be useful predictors of profitability. Therefore, the main objective of this study was to quantify the genetic associations between muscular, skeletal and functional linear type traits with both live-weight and price on animals. A total of 132,376 live-weight and animal price records were available, as well as linear type trait data from 35,669 animals aged between 180 and 364 days. Phenotypic and genetic (co)variance components for animal price, live-weight and linear type traits, were estimated using linear animal mixed models. In general greater prices were paid for taller, deeper and wider animals as evidenced by the moderate genetic correlations between the skeletal and muscularity type traits and animal price. Live-weight, along with 15 type traits, explained 86% of the genetic variation in price; the 15 linear type traits alone explained 73% of the genetic variation in animal price while four selected type traits explained 49% of the genetic variation. Results from this study show that some linear type traits are strongly genetically correlated with both animal live-weight and price and these can be used to enhance the accuracy of selection for animal price and live-weight especially where such data may be limited.
Genetic and environmental impact on auction prices for Holstein cows
2009, Livestock Science
The aim of the present analysis was to determine the impact of a variety of traits and effects (i.e. production, type, health, management effects, pedigree information) on prices of Holstein cows sold at auction, and to estimate genetic (co)variance components between type traits and auction price. Results were used to derive economic weights for type traits. Data of 1565 cows in first parity were collected at six monthly auction sales from August 2005 through January 2006. Seventeen linear type traits and body condition (scale 1 to 9), and four type composites (dairy character, body, feet and legs, and udder; scale 65 to 88) were scored by two classifiers in the auction hall before cows were sold. Analysis of variance revealed a highly significant impact (P < 0.001) of auction date, test day milk yield, stage of lactation, origin of sire, and miscellaneous defects on auction price. The most expensive cows were sold in August, they were from foreign proven sires, they had a high level of test day milk yield, and they were free from defects related to udder, feet and legs, or milkability. The feet and leg, udder, and body composite also had a significant effect on the price (P < 0.001), with higher scores being associated with higher prices. The opposite association was found for dairy character (P < 0.01). Utilizing results from regression analysis, economic weights per genetic standard deviation were highest for linear scored rear udder height (1.23 €), front teat placement (0.97 €), and strength (0.80 €), but were negative for dairy character (− 0.69 €). Genetic parameters for linear type traits scored at the auction date were consistent with literature reports. Heritability for auction price was 0.27, and auction price was genetically positively related to the feet and leg (0.55), udder (0.55), and body composite (0.21). A relative breeding value for auction price was estimated for 27 influential sires, and correlated with official indices for production, conformation, somatic cell count, functional herd life, fertility, and the total net merit index. Correlations were 0.15, 0.21, 0.11, 0.03, 0.05, and 0.19, respectively. Auction price in combination with type scores and information related to farm management provide valuable information for genetic analysis in dairy cattle, and results can be used to increase dairy cow profitability.
Genetic analysis of locomotion and associated conformation traits of Holstein-Friesian dairy cows managed in different housing systems
2008, Journal of Dairy Science
The objectives of this study were to determine the influence of housing on lameness-related linear and composite traits, and to estimate heritabilities of the traits and correlations among them. Data comprised 156,770 national type evaluation records of pedigreed first-lactation Holstein-Friesian cows that calved from 2000 through 2006 and were classified in different housing systems—cubicles, straw yards, slatted or loafing yards, and on pasture. Locomotion score (LOCO), rear leg, side view (RLS), foot angle (FA), bone quality (BO-NEQ), legs and feet (L&F), and mammary composite (MAMM) were the traits measured. Data were analyzed by REML, using an animal model. In general, cows in grazing systems had better locomotion, straighter RLS, steeper FA, flat and more refined bones, better L&F, and better mammary systems compared with cows housed in other systems. Estimates of heritability ranged from 0.11 for LOCO to 0.31 for MAMM. Bone quality had the highest heritability (0.23) of the traits associated with L&F. Genetic associations between BO-NEQ and LOCO, L&F, and MAMM were moderate to high (0.30 to 0.50), but estimates between BONEQ and RLS and FA were not significantly different from zero. Locomotion score had a very high genetic (0.98) and phenotypic (0.78) correlation with L&F, indicating that both traits are genetically the same. On the basis of the genetic parameters, including BONEQ in a selection index as a predictor of longevity is promising, but further information on its association with longevity is required.
Body trait profiles in holstein-friesians modeled using random regression
2005, Journal of Dairy Science
Legendre polynomial and cubic spline functions were used in random regression models to model the change in body traits over the course of the first lactation for daughters of 954 sires. Both functions estimated similar genetic variances for d 50 to 250 across lactation for the majority of traits. The heritability of the traits was similar to other studies using univariate models as well as random regression models. There was little difference between the 2 functions in their predictive power for each of the body type traits, as measured by the absolute difference between the predicted and actual type traits and the proportion of the total phenotypic variance explained by the model. Overall, the Legendre polynomial appeared to model these traits slightly better. Plots of the fixed curves and daily sire solutions obtained from the random regression models showed that there were differences in how the traits and sires changed across lactation. The daily sire solutions were then used to predict differences in liveweight of sires’ daughters across first lactation and showed that the daughters of some sires grew faster during first lactation than others. The spatial differences in the body traits that are displayed by this study could be an important indicator of the physical and biological changes that cows are undergoing in their first lactation. Information from these sire profiles could be harnessed to indicate production and functional traits later in life.

View all citing articles on Scopus

^∗: Present address: University of Oxford, Agricultural Science Building, South Parks Road, Oxford OX1 3PF, Gt. Britain.

View full text

Abstract

Résumé

Zusammenfassung

Livest. Prod. Sci.

J. Dairy Sci.

J. Dairy Sci.

J. Dairy Sci.

J. Dairy Sci.

Type classification of dairy cattle: a review

Anim. Breed. Abstr.

Introduction to linear type assessment

Empirical Bayes estimation of parameters for n polygenic binary traits

Genet., Sel., Evol.

On selection criteria and estimation of parameters when the variance is heterogeneous

Theor. Appl. Genet.