HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Crews, D. H. Articles by Kemp, R. A. Search for Related Content PubMed PubMed Citation Articles by Crews, D. H., Jr. Articles by Kemp, R. A.

Genetic evaluation of carcass yield using ultrasound measures on young replacement beef cattle¹

Agriculture and Agri-Food Canada Research Centre, Lethbridge, Alberta T1J 4B1 Canada

² Correspondence:
5403 1st Avenue South (phone: 403-317-2288; fax: 403-382-3156; E-mail:
dcrews{at}em.agr.ca).

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

 
Live weight and ultrasound measures of fat thickness and longissimus muscle area were available on 404 yearling bulls and 514 heifers, and carcass measures of weight, longissimus muscle area, and fat thickness were available on 235 steers. Breeding values were initially estimated for carcass weight, longissimus muscle area, and fat thickness using only steer carcass data. Breeding values were also estimated for weight and ultrasound muscle area and fat thickness using live animal data from bulls and heifers, with traits considered sex-specific. The combination of live animal and carcass data were also used to estimate breeding values in a full animal model. Breeding values from the carcass model were less accurate and distributed more closely around zero than those from the live data model, which could at least partially be explained by differences in relative amounts of data and in phenotypic mean and heritability. Adding live animal data to evaluation models increased the average accuracy of carcass trait breeding values 91, 75, and 51% for carcass weight, longissimus muscle area, and fat thickness, respectively. Rank correlations between breeding values estimated with carcass vs live animal data were low to moderate, ranging from 0.16 to 0.43. Significant rank changes were noted when breeding values for similar traits were estimated exclusively with live animal vs carcass data. Carcass trait breeding values estimated with both live animal and carcass data were most accurate, and rank correlations reflected the relative contribution of carcass data and their live animal indicators. The addition of live animal data to genetic evaluation of carcass traits resulted in the most significant carcass trait breeding value accuracy increases for young replacements that had not yet produced progeny with carcass data.

Key Words: Beef Cattle • Carcass Composition • Genetic Analysis • Ultrasound

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

 
Real time ultrasound (RTU) technology can be used as a supplement to carcass progeny tests for genetic evaluation of carcass traits (Wilson, 1992; Bertrand et al., 2000). In addition to the drawbacks of cost and time, organized carcass progeny tests may also be limited by only providing accurate evaluations on a biased sample of a few popular sires, and little genetic information on females. The addition of live animal data, including RTU, to carcass evaluation programs provides the opportunity to not only evaluate a larger and more random sample of bulls within a population, but also to evaluate the genetic potential of females.

Carcass traits are generally assumed to have moderate to high heritability (Koots et al., 1994a) and genetic correlations with growth and other live animal traits (Koots et al., 1994b; Crews and Kemp, 1999). Recent studies have shown than RTU traits in bulls and heifers have positive genetic correlations with corresponding carcass traits in slaughter animals (Moser et al., 1998; Reverter et al., 2000; Devitt and Wilton, 2001). However, models that treat RTU and carcass traits as sex-specific (Crews and Kemp, 2001) account for genetic correlations among sexes and between live animal and carcass traits that are less than one.

Several beef breed associations have carcass databases generated from performance programs and organized progeny tests (Bertrand et al., 2000). Some have also begun collection of RTU data to further facilitate genetic evaluation of carcass traits. This study was conducted to investigate the potential for RTU data to augment evaluation of carcass yield by comparing EBV produced from up to three different genetic evaluation systems, including the case in which only carcass data were available, in which only live animal data were available from bulls and heifers, and in which both types of data were available.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

 
Data.

In each of three cow-calf cycle years (1994 to 1996 breeding, 1995 to 1997 calving, 1996 to 1998 harvest), approximately 375 composite (0.25 Charolais, 0.25 Simmental, 0.44 British [Angus, Hereford, Shorthorn], 0.06 Limousin) inter se-mated cows produced calves at the Onefour Research Substation near Manyberries, Alberta. Development of this stable composite breed was described by Mwansa et al. (2000). Calves were identified at birth (February to May) and remained with the cow until weaning in October, when the average calf age was 200 d. A sample of 60 bull calves were castrated each year at weaning.

Remaining bulls and heifers were placed in drylot and fed a growing ration (200 d) that resulted in ADG of 0.95 and 0.80 kg/d, respectively. Live weight and RTU measures of longissimus muscle area and fat thickness were taken at approximately 1 yr of age (371 ± 14 d). Details regarding collection of RTU measurements were described by Crews and Kemp (2001) and followed Beef Improvement Federation recommendations (BIF, 1996).

Steers from each calving year were transported following weaning to the Lethbridge Research Centre feedlot. Steers (n = 60) produced from the same herd that were born in 1994 and managed similarly were also included in this study. Steers were fed a growing ration (150 d, 1.13 kg ADG) and then a finishing ration (90 to 120 d, 1.34 kg ADG) until designated for slaughter when live weight and RTU fat thickness reached minimums of 500 kg and 7 to 10 mm, respectively. Steers (459 ± 21 d) were transported to the Lacombe Research Centre and processed according to standard practices. Carcass data available for this study included hot carcass weight, longissimus muscle area, and fat thickness and were collected approximately 24 h postmortem by a certified beef grader.

The final data set (n = 1,153) consisted of bulls (n = 404) and heifers (n = 514) with live animal data and steers (n = 235) with carcass data. Collection of yearling weights on bulls and heifers is standard practice at Onefour; therefore, records for this trait were available for a larger number (n = 2,698) of animals in this population. Complete parentage information was assembled, including a minimum of four ancestral generations for each animal, beginning with those animals for which RTU or carcass data were available. No animal had both live and carcass measurements. A total of 70 sires had progeny with yearling RTU measurements (13.11 progeny per sire), and 74 sires had steer progeny with carcass data (3.18 progeny per sire). The 70 sires with progeny with yearling RTU data also had steer progeny with carcass data.

Breeding Value Models.

A series of animal models were fit to estimate breeding values and their associated accuracy using software tools included in the Animal Breeder’s Tool Kit (B. L. Golden, personal communication). Fixed effects for yearling weight included year of birth x sex contemporary groups, age of dam (2, 3, 4, 5 yr) classes, plus the linear effects of age at measurement. Ultrasound and carcass trait models included fixed effects for year of birth and the linear effect of age at measurement. Only direct animal genetic effects were included in the random portion of evaluation models.

The full (F) evaluation model included both live and carcass measurements. Using matrix notation, F can be represented by

where y_B, y_H, and y_S = subvectors of observations on live bulls, live heifers, and steer carcasses, respectively; X_B, X_H, and X_S = known incidence matrices relating live bull, live heifer, and steer carcass observations to their respective fixed effects; b_B, b_H, and b_S = subvectors of live bull, live heifer, and steer carcass fixed effects; Z_B, Z_H, and Z_S = known incidence matrices relating live bull, live heifer, and steer carcass observations to their respective random effects; u_B, u_H, and u_S = subvectors of live bull, live heifer, and steer carcass random additive genetic animal effects; and e_B, e_H, and e_S = subvectors of live bull, live heifer, and steer carcass random residuals. Random components of this model had null expectation, and the following assumed (co)variance structure:

and

where A is the additive numerator relationship matrix, I is the identity matrix of order appropriate to the numbers of observations. Subscripts B, H, and S denote live bull, live heifer, and steer carcass components, respectively. The A matrix was constructed for 4,356 animals, including 226 base animals without parentage information or data. Diagonal elements of A were greater than one for 698 animals, whose average inbreeding coefficient was 0.03. Additive genetic (_u²) and residual (_e²) variances and genetic covariances were those estimated by Crews and Kemp (2001). Because no animal had both live animal and carcass measurements, residual covariances were constrained to zero.

The full model for weight was slightly different because yearling live weights were assumed to be genetically equivalent between bulls and heifers (r_g = 0.98) but separate from carcass weight of steers (Crews and Kemp, 2001). Therefore, model F for weight was a bivariate model, whereas F was a three-trait model for longissimus muscle area and fat thickness as shown above. Reduced models including only steer carcass (C) or only live yearling bull and heifer (L) measurements were also fit. Matrix representation of C and L by reduction of F is straightforward. Table 1 summarizes the data used and EBV estimated with each model for each trait. Accuracy of EBV was calculated by the software as described by the Beef Improvement Federation ( BIF, 1996). Models for weight, longissimus muscle area, and fat thickness were fit separately because of computational limitations involved with high-order multiple-trait models; however, it is recognized that components of carcass yield are correlated (Koots et al., 1994b; Crews and Kemp, 1999).

To compare EBV estimated from F, L, and C, a reference set of animals was designated including all animals used to compute EBV minus base parents (i.e., without parentage information or data) and steers. Solutions for animal genetic effects were estimated for the animals not in the reference set; however, base parents (n = 226) were far removed from the data (three to four generations) and it was of interest to compare EBV only among potential replacements.

When two EBV are estimated with perfect accuracy and based on independent sources of information, the expected value of the simple correlation between those two EBV is equal to the genetic correlation between the two traits. With less than perfectly estimated EBV, the expected simple correlation is a function of accuracy and genetic correlation (L. D. Van Vleck, personal communication). In general, rank correlations do not have this expectation. However, rank correlations are preferable when EBV based on (sub)sets of related data are compared.

Summary statistics and rank correlations among EBV were computed using SAS (SAS Inst. Inc., Cary, NC). Three accuracy class categories were defined on the basis of EBV accuracies from model C. Animals with EBV from C with accuracy 0.30 were classified as low accuracy. Bulls and heifers that had not produced steer progeny with carcass data would likely have low accuracy carcass EBV. Animals with EBV from C with accuracy 0.60 were classified as high accuracy. The high accuracy category contained the fewest numbers of animals. The remaining animals (0.30 < accuracy < 0.60) were classified as moderate accuracy.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

 
Table 2 contains a summary of live animal (i.e., yearling bull and heifer) and carcass (i.e., steer) phenotypic and genetic parameter estimates. Summary statistics and genetic parameters used in these analyses were those reported by Crews and Kemp (2001), based on this population. Numerically, yearling bulls had less fat than heifer contemporaries (4.07 vs 5.62 mm) and larger muscle area (69.07 vs 58.62 cm²). The average steer carcass weighed 300 kg with fat thickness of 8.76 mm and muscle area of 82.80 cm². Heritability estimates for the live animal and carcass traits were moderate to high (0.38 to 0.69) and generally within the ranges published in the literature (Koots et al., 1994a; Reverter et al., 2000). One possible exception was yearling weight, for which the heritability estimate (0.69) was higher than some published estimates (Koots et al., 1994a; Moser et al., 1998) and higher than parameters used in national cattle evaluations for some breeds (E. J. Pollak, personal communication; J. K. Bertrand, personal communication). Relatively few numbers of observations in this population led to parameter estimates with moderately large standard errors. Robinson et al. (1993) reported a mean heritability of 0.46 ± 0.04 for yearling weight based on five populations, but the estimates ranged from 0.31 to 0.64. Genetic correlations of live animal measures from bulls and heifers with corresponding steer carcass measures are also summarized in Table 2. Most of the genetic correlations are similar to those reported in recent literature (Moser et al., 1998; Reverter et al., 2000; Devitt and Wilton, 2001), with the exception of the genetic correlation between yearling bull RTU and steer carcass fat thickness (0.23). Moser et al. (1998) reported 0.69 for this correlation in Brangus, which was supported by estimates of 0.79 and 0.87 in Angus and Hereford bulls, respectively, reported by Reverter et al. (2000). The lower correlation used here may be attributable to breed differences (i.e., previous estimates are largely based on cattle with little or no continental European breed composition) and(or) to the lower mean RTU fat thickness (4.07 mm) of bulls in this study.

Weight Breeding Values.

Four EBV were produced for measures of weight (Table 1). Because yearling weights of bulls and heifers were considered genetically equivalent (Crews and Kemp, 2001), separate EBV for bulls and heifers for this trait were not produced. Table 3 contains summary statistics for weight trait EBV and their accuracies. The range of EBV for yearling weight was considerably larger than the range of carcass weight EBV, regardless of data used in their computation. This was likely due to the higher mean, larger variance, and higher heritability of yearling weight. Due to fewer steer carcass weight observations and the lower heritability of carcass weight, model C was, on average, least accurate for genetic evaluation of weight. However, evaluation of carcass weight using both steer carcass and yearling bull and heifer data produced EBV with the highest mean accuracy (0.61).

Muscle Area Breeding Values.

Crews and Kemp (2001) reported genetic correlations less than one both between RTU muscle area of heifers vs bulls and between RTU muscle area of potential replacements with carcass muscle area of slaughter progeny. Table 4 summarizes EBV and accuracies for six muscle area EBV. Model L included yearling RTU muscle area data from bulls and heifers, modeled as separate but correlated traits. The range of yearling bull RTU muscle area EBV was approximately 50% larger than that for yearling heifer RTU muscle area EBV, most likely due to larger mean, variance, and heritability of bull measures. Yearling heifer RTU muscle area EBV were generally more accurate, due probably to more animals having records. Yearling RTU muscle area EBV were also estimated with model F, which included both live and carcass measurements. The addition of carcass data with model F resulted in bull and heifer RTU muscle area EBV that were, on average, more accurate for bull measurements but no change in accuracy for heifer measurements. In the case of both yearling bull and heifer RTU muscle area, the range of EBV was similar (less than 1% difference in range across gender), regardless of the addition of steer carcass muscle area data (Model F vs L). Similar to weight models, model C contained the fewest number of records and therefore generally produced the least accurate evaluation. Also similar to the results noted for weight, carcass muscle area EBV from model F included the largest amount of data and the highest mean accuracy.

Fat Thickness Breeding Values.

Table 5 summarizes fat thickness EBV and accuracies and contains rank correlations among EBV estimated from the three models. Similar to results noted for weight and muscle area, models would generally be ranked F > L > C with respect to mean EBV accuracy. This ranking was probably due to the relative numbers of observations included in the analyses. Even though carcass fat thickness had higher variance than RTU fat thickness measurements, carcass fat thickness EBV from model C were least variable, probably due in part to the lower heritability estimate (0.38) for carcass vs RTU fat thickness (0.44 to 0.50).

Weight Accuracy Categories.

Table 6 provides rank correlations among the four weight EBV estimated with models L, C, and F across accuracy categories. Accuracy category was designated on the basis of accuracy values for carcass weight EBV from model C. Model C utilized the least amount of data (i.e., only carcass data from steers), and therefore C was the least accurate model on average, and numbers of potential replacements in each accuracy category were different. The 2,588 bulls and heifers in the low accuracy category had a mean carcass weight EBV accuracy from model C of 0.29. Weight EBV estimated using either model L or model F were more accurate. Similar to results noted previously, rank correlations of carcass weight EBV from model C were only moderate with weight-related EBV from the other models. Model F carcass weight EBV had nearly perfect (0.98) rank correlations with yearling weight EBV from model L and from model F. A similar trend was noted for moderate accuracy animals in category two. Models L and F provided more accurate genetic evaluations on average than model C, due primarily to the increased amount of data. Carcass weight EBV from model F had nearly perfect rank correlations with yearling weight EBV from both models L and F; however, carcass weight EBV from model C had lower rank correlations with weight EBV from other models. An almost identical trend was also noted for high accuracy animals in category three. The reduction in rank correlation of carcass weight EBV from model C was smaller for category three compared to categories one and two.

Muscle Area Accuracy Categories.

Results of the analysis of muscle area accuracy categories (Table 7) were similar to those for weight, and the same number of animals were categorized into each accuracy category. As expected, the average EBV accuracy increased with category and when EBV were estimated from models including higher numbers of observations. In category one, in which carcass muscle area EBV had lower accuracy, muscle area EBV from model F had higher rank correlations with muscle area EBV from model L (0.82 to 0.86) than with carcass muscle area EBV from model C. As accuracy category increased and carcass muscle area EBV were more accurate with model C, model F carcass muscle area EBV had more similar (category two) or higher (category three) rank correlations with carcass muscle area EBV from model C vs muscle area EBV from model L. As expected, when carcass data available are sufficient to provide high accuracy evaluations, the contribution of live animal data to accuracy of evaluation through genetic correlation is reduced. Likewise, for young bulls and heifers prior to their production of progeny with carcass data, carcass trait evaluations will have lower accuracy, and those evaluations will be estimated largely through genetic correlation and relationships. The increases in accuracy for these animals will be generated entirely through addition of live animal data (e.g., yearling weights, RTU data), making the proper modeling of this data important.

Table 8 summarizes mean accuracies and rank correlations among fat thickness EBV across accuracy class. Fat thickness EBV from model C had average accuracies of 0.29, 0.39, and 0.66 in the low, moderate, and high accuracy categories, respectively, similar to the results noted for weight and muscle area. Also similar to the results noted for muscle area, fat thickness EBV from model C had lower correlations with RTU fat thickness EBV from model L. As previously noted, the lower association between EBV estimated using different types of data is indicative of major rank changes depending on the type of data used to estimate genetic evaluations. Corresponding EBV from models L and F had nearly perfect rank correlations, indicating that the addition of carcass data to a live animal data evaluation increased mean accuracy but did not significantly change relative animal ranks. As accuracy of carcass fat thickness EBV (model C) increased with accuracy category, the rank correlation with carcass fat thickness EBV from model F increased from 0.71 to 0.86.

	Implications

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

	Footnotes

 
¹ AAFC-LRC contribution no. 38701017. These data were collected with funding support from the Canadian Beef Industry Development Fund for project 95H022. The authors gratefully acknowledge the contributions of the staff at the Onefour Research Substation and Lethbridge Research Centre Feedlot for cattle management and data collection, as well as N. Shannon and R. Crews for ultrasound image collection and analysis. The contributions of the Lacombe Research Centre, especially J. Aalhus and W. Robertson, are also acknowledged.

Received for publication June 19, 2001. Accepted for publication March 8, 2002.

	Literature Cited

Top Abstract Introduction Materials and Methods Results and Discussion Implications Literature Cited

 


BIF. 1996. Guidelines for Uniform Beef Improvement Programs. 7th ed. Kansas State Univ., Colby.

Bertrand, J. K., D. W. Moser, and W. O. Herring. 2000. Beef genetic evaluation programs for carcass traits: current situation and future possibilities. J. Anim. Sci. 78(Suppl. 1):57 (Abstr.).

Crews, D. H., Jr., and R. A. Kemp. 1999. Contributions of preweaning growth information and maternal effects for prediction of carcass trait breeding values among crossbred beef cattle. Can. J. Anim. Sci. 79:17–25.

Crews, D. H., Jr., and R. A. Kemp. 2001. Genetic parameters for ultrasound and carcass measures of yield and quality among replacement and slaughter beef cattle. J. Anim. Sci. 79:3008–3020.[Abstract/Free Full Text]

Devitt, C. J. B., and J. W. Wilton. 2001. Genetic correlation estimates between ultrasound measurements on yearling bulls and carcass measurements on finished steers. J. Anim. Sci. 79:2790–2797.[Abstract/Free Full Text]

Golden, B. L. 2001. Genetic prediction for time to finish end points in beef cattle. J. Anim. Sci. 79 (Suppl. 1):99 (Abstr.).

Golden, B. L., D. J. Garrick, S. Newman, and R. M. Enns. 2000. A framework for the next generation of EPDs. In: Proc. 32nd Beef Improv. Fed. Annu. Res. Symp. and Mtg., Wichita, KS.

Koots, K. R., J. P. Gibson, C. Smith, and J. W. Wilton. 1994a. Analyses of published genetic parameter estimates for beef production traits. 1. Heritability. Anim. Breed. Abstr. 63:309–338.

Koots, K. R., J. P. Gibson, and J. W. Wilton. 1994b. Analyses of published genetic parameter estimates for beef production traits. 2. Phenotypic and genetic correlations. Anim. Breed. Abstr. 62:825–853.

Moser, D. W., J. K. Bertrand, I. Misztal, L. A. Kriese, and L. L. Benyshek. 1998. Genetic parameter estimates for carcass and yearling ultrasound measurements in Brangus cattle. J. Anim. Sci. 76:2542–2548.[Abstract/Free Full Text]

Mwansa, P. B., R. A. Kemp, D. H. Crews, Jr., J. P. Kastelic, D. R. C. Bailey, and G. H. Coulter. 2000. Comparison of models for genetic evaluation of scrotal circumference in crossbred bulls. J. Anim. Sci. 78:275–282.[Abstract/Free Full Text]

Reverter, A., D. J. Johnston, H.-U. Graser, M. L. Wolcott, and W. H. Upton. 2000. Genetic analysis of live-animal ultrasound and abattoir carcass traits in Australian Angus and Hereford cattle. J. Anim. Sci. 78:1786–1795.[Abstract/Free Full Text]

Robinson, D. L., K. Hammond, and A. C. MacDonald. 1993. Live animal measurements of carcass traits: Estimation of genetic parameters for beef cattle. J. Anim. Sci. 71:1128–1135.[Abstract]

Wilson, D. E. 1992. Application of ultrasound for genetic improvement. J. Anim. Sci. 70:973–983.[Abstract]

This article has been cited by other articles:

				D. H. Crews Jr., E. J. Pollak, and R. L. Quaas Evaluation of Simmental carcass EPD estimated using live and carcass data J Anim Sci, March 1, 2004; 82(3): 661 - 667. [Abstract] [Full Text] [PDF]

				D. H. Crews Jr., E. J. Pollak, R. L. Weaber, R. L. Quaas, and R. J. Lipsey Genetic parameters for carcass traits and their live animal indicators in Simmental cattle J Anim Sci, June 1, 2003; 81(6): 1427 - 1433. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Crews, D. H. Articles by Kemp, R. A. Search for Related Content PubMed PubMed Citation Articles by Crews, D. H., Jr. Articles by Kemp, R. A.