GENETIC AND ENVIRONMENTAL EFFECTS ON ...

3 downloads 0 Views 941KB Size Report
Approximately 4,400 crossbred lambs from Southdown sires and Romney ewes were slaughtered at approximately 18, 23, and 28 wk of age over a 16-yr period.
Genetic and environmental effects on carcass characteristics of Southdown x Romney lambs: II. Genetic and phenotypic variation G. L. Bennett, D. L. Johnson, A. H. Kirton and A. H. Carter J ANIM SCI 1991, 69:1864-1874.

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://jas.fass.org/content/69/5/1864

www.asas.org

Downloaded from jas.fass.org by guest on July 10, 2011

GENETIC AND ENVIRONMENTAL EFFECTS ON CARCASS CHARACTERISTICS OF SOUTHDOWN x ROMNEY LAMBS: II. GENETIC AND PHENOTYPIC VARIATION G. L. Bennett1, D. L. Johnson, A. H. Kirton and A. H. Cartes Ruakura Animal Research Station, Private Bag, Hamilton, New Zealand ABSTRACT

Approximately 4,400 crossbred lambs from Southdown sires and Romney ewes were slaughtered at approximately 18, 23, and 28 wk of age over a 16-yr period. Live weights, carcass measurements, and chemical percentages were analyzed to estimate genetic and phenotypic parameters. Heritabilities of postweaning weights and gains were about .20. Heritabilities of fat and water percentages were about .35 adjusted for age. Heritability of kidney fat percentage was .53. Heritabilities of fat depth and muscle measurements ranged from .21 to .37. Crutch depth (h2= .73) and cannon bone length (h2 = .74) were the most highly heritable carcass measurements. The genetic correlation &wen carcass fat and fatfree weight was .47 when lambs were slaughtered at a constant age. Fat-free weight was nearly uncorrelated with percentages of fat, water, and protein when lambs were slaughtered at the same age. Carcass measurements increased accuracy of selection for fatfree weight at a constant age very little compared with using only carcass weight. However, this does not mean that additional measurements are useless. The addition of carcass measurements to the selection criteria would result in correlated responses in chemical composition that more closely resembled direct selection for fat-free weight. Carcass weight would be of little value when used by itself to reduce fat weight adjusted for carcass weight. Direct measurement of carcass composition resulted in 1.6 to 2.6 times more predicted response for reduced fat weight than any combination of carcass weight and one fat depth measurement. Key Words: Carcass Composition, Sheep, Genetic Correlation, Phenotypic Correlation, Heritability, Genetic Variance I. Anim. Sci. 1991. 69:186&1874

introduction

genetic parameters for chemical or separable Improvement of carcass composition by tissue composition and related measurements of lamb carcasses has been done in larger selection depends ultimately on its genetic studies by Botkin et al. (1969, 1971) and Wolf variance. The rate and cost of improvement et al. (1981). These studies indicated genetic depends on the heritabilities of composition standard deviations of about 2% for fat and and indicator traits and the phenotypic and lean in animals slaughtered at the same live genetic correlations among them. Estimation of weight. Furthermore, lean and fat were found to be correlated with growth rates and other easily obtained carcass measurements. The purpose of this paper is to estimate 'Present address: Roman L. Hruska U.S. Meat Animal heritabilities and genetic correlations among Res. Center, A R S , USDA, P.O.Box 166, Clay Center,NE growth and carcass traits in Southdown x 68933-0166. Romney lambs. These estimates are used to 2DeCeaSed. determine the usefulness of indirect measureRexeived January 2, 1990. Accepted November 6, 1990. ments for different selection objectives.

1864

Downloaded from jas.fass.org by guest on July 10, 2011

GENEnC PARAMETERS OF LAMB CARCASSES

Materialsand Methods

Experimental Design Lambs and traits analyzed were described previously (Bennett et al., 1991). Sire selection procedures are now given in more detail because they affect the genetic interpretation of sire variance components. Initially, purebred Southdown rams were sampled from industry flocks in New Zealand. This industry sampling began with the 1963 mating and continued through the 1972 mating. However, rams from an experimental Southdown flock at Ruakura made up the majority of rams used from the 1965 to 1972 mating and all rams used from 1973 through 1978.The experimental flock at Rualcura was begun with the 1965 mating using sires originally sampled for progeny-test matings with Romney ewes. The experimental flock was closed to further introduction of industry sires in 1965. The experimental flock consisted of three lines. Sires from one line were selected for live weight at 6 mo of age and sires from a second line were selected for live weight gain from 3 to 6 mo of age. Sires from the third line were initially selected on growth performance and subsequently selected for growth and carcass weight on the basis of progeny-test results. Each line used three sires each year. Sires were replaced by a son if possible. Some rams were used in more than one year. Only data recorded during the first year of use were included in the analyses. Rams from the experimental flocks tended to be those that were selected on their line criterion, but some more poorly performing rams were used to validate selection prccedures. From 1973 to 1978,rams from the third line were used exclusively.

1865

homogeneous across years. The sums of squares and cross products obtained from each year were combined to produce overall estimates of within- and between-sire variance and covariance components. A method that accounted for the expected Gamma distribution of the estimated variance and covariance components and for their multivariate relationship was developed and used. The loglikelihood of the multivariate distribution was determined and then maximized by setting the first derivatives to zero. Second derivatives were used to obtain numerical solutions by the method of scoring. Although not a strictly balanced design, the sire and residual sums of squares and cross products for each year were assumed to have independent central Wishart distributions. For year j, the between- and within-sire sums of squares and CTOSS products, Bj and Wj, respectively, are distributed as

w: -B.

Wp(dj,k&+RJ Wp(ej,QJ

where dj = degrees of freedom for sires, kj = expected sire component coefficient, ej = residual degrees of freedom, p = number of traits or dimension of sums of squares and products matrix, and R,,Re are the sire and error variancecovariance matrices that are to be estimated. The log-likelihood (L) is then given by

+ (dj - p - 1)logJB) - tr(R7'B J J.)/2) + C(-.jlogJRJn j

+