J. Anim. Sci. 2004. 82:3366-3372
© 2004 American Society of Animal Science
The relationship of average backfat thickness of feedlot steers to performance and relative efficiency of fat and protein retention
J. R. Brethour1
Kansas State University Agricultural Research Center, Hays 67601
 |
Abstract
|
|---|
Selection for growth and improved carcass merit has resulted in cattle that are variable in composition of gain during the finishing phase. This study assessed the relative performance among cattle with different levels of initial backfat thickness. It also exploited the ability to track carcass composition in the live animal with ultrasound estimates of backfat and marbling. A procedure was developed to partition and estimate relative efficiency of fat and protein gain. The trial periods were the last 43 or 50 d before slaughter and included 10 pens (average of 27 animals per pen) that ranged in average backfat thickness from 6.3 to 13.1 mm. There was no correlation (r2 = 0.0026) between average backfat thickness and G:F (g/kg of DMI). Correlations between average backfat thickness and ADG or DMI were also nearly zero (r2 = 0.0007 and 0.0042, respectively). Fat deposition from NEg was 3.98 times more efficient than protein deposition. Carcass backfat thickness was a poor predictor of carcass marbling score (r2 = 0.083), even though backfat thickness was an important predictor of the percentage of empty body fat (r2 = 0.807). The results indicate that a measure of backfat thickness on the live animal during the finishing phase is not an effective predictor of future feed efficiency. They also confirm that protein accretion is energetically expensive, and that using a single coefficient for predicting gain from NEg is valid regardless of whether gain is predominately muscle or fat. These data document that there is little relationship between body composition and marbling score, which is contrary to models that assume a USDA quality grade target at a specified percent fat end point.
Key Words: Beef Cattle • Body Composition • Fat • Protein • Ultrasound
|
Introduction
|
|---|
Changes in the cattle feeding industry, such as grade and yield pricing and individual tracking technology, have generated interest in precision marketing individual animals. There is an optimal slaughter date (days on feed) for each animal that maximizes profit, and there is an average loss in return of approximately $1 for each day that actual slaughter differs from the optimum (Koontz et al., 2000
).
Models to project optimal slaughter date estimate when incremental production costs equal incremental increases in animal value. The latter are comprised of increases in carcass weight and carcass value, which is determined by the schedule of premiums and discounts specific to a price grid. The primary incremental cost is feed; thus, estimates of future feed efficiency are crucial to an effective model.
Feed conversion has been estimated from formulas stating feed requirements per unit of live gain (NRC, 1996). This method assumes that efficiency decreases as more fat and less protein constitutes gain because of the higher caloric content and lower moisture content of fat tissue. Composition of gain is regarded as a function of animal BW and ADG in calculating energy requirements; however, ultrasound provides a rapid and economical way to estimate body composition. In addition, ultrasound backfat measures made upstream in the feeding phase accurately project future backfat thickness up to and including slaughter (Brethour, 2000).
This study was conducted to estimate feed efficiency as a function of average backfat thickness determined with ultrasound at the beginning of the trial and of carcass backfat at slaughter. Although it is known that the energy cost of protein gain is greater than fat gain, there are few reports that quantify this difference. The opportunity to use ultrasound to estimate carcass composition in the live animal was exploited to develop a procedure that partitioned retained energy between fat and protein and calculated the relative efficiency of fat and protein gain.
|
Materials and Methods
|
|---|
The design for this experiment consisted of two trials, each with five pens averaging 27 steers per pen. Before the start of the trial, backfat thickness was measured with an Aloka 210 (Wallingford, CT) ultrasound system using a 5-MHz transducer array (UST-5813N-5). This probe provides much greater accuracy and precision than 3.5-MHz probes because estimates can be made to a fraction of a millimeter. Backfat measurements were taken just before the beginning of the trials with a transverse orientation between the 12th and 13th ribs approximately 10 cm distal from the back. Cattle were ranked by backfat thickness and assigned to treatment groups. Animals that were predicted to have more than 18 mm of backfat (equivalent to USDA Yield Grade 4) before the completion of the trials were omitted. In an attempt to have average initial starting weights as equal as possible, the heaviest cattle with low backfat thickness were chosen to represent the thinnest group in each trial. This research was conducted from August to October 2002, a period when there was little environmental stress. The protocol was approved by the Kansas State University Institutional Animal Care and Use Committee.
Cattle used in this study were all black (at least 50% Angus), with no indications of Bos indicus; however, exact parentage was not known. They had been purchased locally the previous fall and raised together in a growing facility before they were acquired for the finishing phase. The pretrial periods on high-concentrate diets for Trials 1 and 2, respectively, were 102 and 94 d. The end-stage experimental periods lasted 43 and 50 d, respectively, and occurred between August 28 and October 23, 2002.
Diets were comprised primarily of finely rolled milo plus sorghum silage, soybean meal, urea, ammonium sulfate, calcium carbonate, and trace mineral premix (Table 1). Calculated energy content (from tabular values) of the diet was 1.35 and 1.98 Mcal/kg of NEg and NEm, respectively. Monensin and tylosin were fed, and cattle were implanted with progesterone/estradiol benzoate (Synovex-S; Fort Dodge Animal Health, Overland Park, KS) at the beginning of the pretrial period, and with trenbolone acetate/estradiol benzoate (Synovex-Plus, Fort Dodge Animal Health) at the start of the trials.
Slaughter was at a commercial facility. Carcass back-fat was measured to the nearest millimeter, and marbling scores were assigned by USDA graders. Bruise trim was estimated and added to carcass weight. There were two animals with trim so excessive that an accurate estimate was impossible, and their data were discarded. Two other cattle were omitted because it was questionable that the carcass corresponded to the live animal. Marbling scores are expressed as 4.0 = Slight00 and 5.0 = Small00.
Cattle were weighed two times when the trials were started. The second weight was taken 3 d after the first weight and after 12 h without feed or water. Weighing order was recorded during both weigh sessions, and regression analysis was used to establish an initial shrunk weight that was independent of weighing time. Because cattle may have been stressed when ultrasound measures were taken just before the trial started, and there seemed to be exceptional shrink when starting weights were taken, a dressing percent of 65 was applied to estimate initial carcass weight. Final weights were taken the day before cattle were slaughtered and included 4% calculated shrink.
Each pen was considered the experimental unit, and a linear model regressed pen averages of dependent variables of interest on mean backfat thickness to test the significance of association with that variable. The two trials were combined and the regression model included trial and average initial weight. All calculations and statistics were performed on a Lotus 1-2-3 spreadsheet (Borland, Scotts Valley, CA).
Variability of gain was high because of the short duration of the trials, which may have caused the significant kurtosis in the gain data. Both kurtosis and skewness were significant for marbling scores. Consequently a nonparametric procedure (Hodges-Lehman estimator, Hollander and Wolfe, 1973) was applied as the best estimate of location (pseudo-median) of gain and marbling for each pen because it decreases the effect of outliers. Pen means were used for the other traits. To remove the effect of differences in dressing percent on gain, final live weight was calculated by dividing carcass weight by 0.64, even though this resulted in a higher SD for ADG than using actual final live weight shrunk 4% (0.46 vs. 0.33 kg, respectively). This difference in SD may have resulted from the inability to estimate individual dressing percents when the trials were initiated.
Initial and final carcass empty body fat was calculated from an equation published by Guiroy et al. (2001) and modified by Tedeschi et al. (2004) to omit LM area (percentage of empty body fat = 14.08796 + 4.7135 fat thickness, cm + 0.01316 hot carcass weight + 0.90855 marbling score). The proportional effects of those predictors were obtained by comparing the partial regression coefficients in standard measure (beta-primes, Snedecor, 1955). A worksheet of calculations to partition retained energy as fat and protein is shown in Table 4. Estimated available NEg was computed from intake and maintenance requirements. Although all cattle were black, the increase in LM area and the decrease in marbling score as average backfat decreased indicated that the thinner groups likely included Continental-breed crosses. An adjustment in NEm was included assuming that the fattest group was 100% Angus and the thinnest group was 50% Continental-breed cross. Then, half the increase in maintenance requirement (9.4%) that Ferrell and Jenkins (1985) reported among Simmental cattle was added proportionally according to average backfat thickness to maintenance requirements. Finally, an estimate of the relative efficiency of fat and protein deposition was obtained by regressing energy gained (Mcal/d) as fat on energy gained (Mcal/d) as protein and NEg. The partial regression coefficient (independent of available energy and trial) is the ratio of incremental increases in fat and protein energy retention and provides an estimate of the relative energy cost of fat and protein gain.
View this table:
[in this window]
[in a new window]
|
Table 4. Calculations to estimate daily energy gained as fat and protein and to estimate energy available for gain (NEg)
|
|
|
Results and Discussion
|
|---|
Average backfat thickness for the 10 groups ranged from 6.3 to 13.1 mm. Although it might have been of interest to have had even fatter groups, the fattest group averaged 16.3 mm at slaughter, the upper limit in modern beef production. There was no relationship (r2 = 0.003) between average backfat and G:F, as shown in Figure 1. Even if the regression coefficient (–0.0108) were decreased 1 SD (0.0792), there would be only a 4% difference in estimated feed efficiency between the poorest and best (leanest) groups. The failure to observe a treatment effect was unexpected because there were substantial changes in the composition of gain associated with the experimental variable, average backfat thickness (Figure 2). It seemed that feed required for a unit of fat tissue gain was the same as that of fat-free tissue, which would be primarily muscle. Average backfat thickness did not affect feed intake, ADG, or dressing percent (Tables 2 and 3). Gains were exceptionally high in both trials, and carcass gain averaged 1.17 kg/d, possibly because there was little environmental stress. Longissimus muscle area was smaller (P < 0.005) and marbling score higher (P < 0.0005) with the progression from thinner to fatter groups (Table 3). This result was probably associated with more Continental-breed genetics among the thinner cattle.
View larger version (8K):
[in this window]
[in a new window]
|
Figure 1. Relationship of G:F (g/kg of DMI) to average backfat thickness, mm (r2 = 0.0007). G:F = 152.1 – 0.108 (± 0.73) average backfat, mm. Standard error of estimate = 4.86.
|
|
View larger version (10K):
[in this window]
[in a new window]
|
Figure 2. Relationship of composition of gain to initial backfat estimate (r2 = 0.88, P < 0.001). Percent empty body gain as fat = 35.69 + 2.152 (± 0.279) initial backfat, mm. Standard error of estimate = 1.51.
|
|
Nutrient requirement systems (CSIRO, 1990; NRC, 1996) assume that energy costs of gain increase as the proportion of gain as fat increases. That is because of the increased caloric value of fat compared with protein (9.385 and 5.539 kcal/g) and the decreased moisture content of fat tissue (10%, Loveday and Dikeman, 1980) compared with fat-free matter (73%, Reid et al., 1955). In addition, those systems assume that composition of gain is a function of BW and ADG. That assumption was probably true with the British breeds used to develop those systems several decades ago, but more recent progress in genetic selection for growth and carcass merit has greatly decreased that relationship. In this study, there were differences in composition of gain, even though average weight was nearly the same among groups.
Using ultrasound in the protocol seemed to be effective in creating groups with varying composition of gain. The r2 values for correlation of initial backfat and marbling with the respective carcass measures were 0.59 and 0.39, respectively. Proportion of gain as fat increased with increases in initial backfat among groups (Figure 2
). That finding agrees with an exponential model (Y = Ae(kt), where Y = future backfat, A = current backfat, t = days, and k = the rate coefficient) of backfat development (Brethour, 2000). In this model, the increase in backfat thickness over a time interval is positively correlated to initial backfat thickness. In Trials 1 and 2, the observed rate coefficients were 0.010 and 0.011, respectively.
Table 4 shows the calculations to partition fat and protein gain and derive energy balances. The three items used to estimate relative efficiency of fat and protein accretion were daily energy gain as fat or protein and calculated NEg. Including the latter item in the regression model adjusted for differences in intake, gain, and presumed differences in breed composition among the 10 groups. The partial regression of energy retained as fat on energy retained as protein measures the incremental changes across the groups and provides an estimate of the relative efficiency of fat and protein accretion (Figure 3). The regression model (fat accretion, Mcal/d = 12.85 – 3.98 [± 1.72] protein accretion) indicates that fat accretion was 3.98 times more energetically efficient than protein, a finding that is very close to a summary of 52 trials by Geay (1984), where the relative efficiency of ME use for fat and protein deposition was 3.75. Rattray and Joyce (1976) reported that efficiency of fat deposition in sheep ranged from 76.7 to 82.5% and the corresponding range for protein deposition was 10.4 to 20.5%. Averaging those numbers results in estimating that efficiency of fat deposition was 5.15 times the efficiency of protein deposition, which is within the range of the estimate in this report.
View larger version (11K):
[in this window]
[in a new window]
|
Figure 3. Ratio of fat accretion to protein accretion, Mcal/d (r2 = 0.40, P < 0.05). Fat accretion, Mcal/d = 12.85 – 3.98 (± 0.53) protein accretion, Mcal/d. Standard error of estimate = 0.53.
|
|
The relative inefficiency of protein accretion is well known (Baldwin et al., 1980). Much of that is caused by the dynamic behavior of protein and concomitant losses associated with protein turnover. Lobley (1992) reported a summary of net N balances in growing lambs that indicated that net retained N was only 23% of absorbed amino acid N. Rattray and Joyce (1976) discussed the importance of protein nutrition and energy efficiency of protein retention, but the excellent performance of animals in these trials would seem to indicate that the diet was satisfactory.
Assuming that fat tissue contains 85% fat (10% water and 5% ash) and muscle contains 22% protein (73% water and 5% ash), and that caloric values of fat and protein are 9.385 and 5.539 kcal/g respectively, an expected relative efficiency of fat and protein retention might be 6.55, where there was an equivalent replacement between fat and muscle gain, as seen in this trial. Much of the discrepancy between that estimate and the relationship of 3.98 determined from regression may have resulted from including a provision for increased maintenance requirements in the leaner cattle. An equivalent model that omits NEg/d is as follows: fat accretion = 13.53 – 4.65 (± 1.93) protein accretion; this model results in a higher estimate of the relative efficiency of fat deposition (4.65). Also, there may be a significant amount of energy that is not protein (e.g., glycogen) in fat-free tissue. There is a possibility that the increase in maintenance requirement used in the calculations for the putative breed composition shift was insufficient. More nervous temperament and greater muscle mass may increase the amount of energy lost in futile cycles among some breeds.
Although pen averages for marbling scores were highly correlated with backfat thickness (r2 = 0.72, df = 8), analysis of individuals resulted in a lower correlation (r2 = 0.10, df = 268). Figure 4 shows the relationship of carcass marbling and estimated percentage of empty body fat at slaughter. When calculated from the regression equation, USDA Choice (marbling score = 5) was equivalent to 28% empty body fat. The NRC (1996) indicated that USDA Choice should be reached at 27% empty body fat; however, the formula (Tedeschi et al., 2004) for estimating percentage of empty body fat includes marbling score, which contributed 33% of the sensitivity of that equation (the other variables in that model, backfat thickness and carcass weight, contributed 56 and 11%, respectively). Percentage of separable carcass fat is highly correlated with backfat measured by ultrasound on the live animal or carcass backfat (Brethour, 1997; Greiner et al., 2003); hence, backfat should be a good single predictor of percentage of empty body fat. When only carcass backfat was correlated with carcass marbling (the partial correlation between carcass weight and marbling was almost zero in a model with backfat as the only other predictor), the r2 value was only 0.08, which was statistically significant (P < 0.01) but which would provide little utility as a predictor (Figure 5).
View larger version (14K):
[in this window]
[in a new window]
|
Figure 4. Regression of marbling score (Slight00 = 4.0; Small00 = 5.0) on percentage of empty body fat (at slaughter; r2 = 0.431, P < 0.01). Marbling score = – 3.99 + 0.313 (± 0.022) percent empty body fat. Standard error of estimate = 0.753.
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Figure 5. Regression of marbling score on carcass back-fat (r2 = 0.083, P < 0.01). Marbling score = 4.47 + 0.088 (± 0.018) backfat, mm. Standard error of estimate = 0.96.
|
|
Retained energy estimated with the NRC (1996) equation averaged 12.77 Mcal/d, which was much higher than the 9.61 Mcal/d average calculated from the formula that estimated empty body fat (Tedeschi et al., 2004). This is because the NRC (1996) formula assumes that the proportion of gain as fat is a function of BW and ADG; however, a single term to describe the efficiency of use of NEg for gain seems justified by these results that indicate muscle gain has the same energy requirement as fat tissue gain. The CSIRO (1990) stated that an alternative to a single term for growth efficiency is not feasible because efficiency of energy used for protein gain is highly variable and is affected by physiological state, feeding level, and feed quality. Efficiency of protein retention might be increased by the calpain/calpastatin system in Bos indicus cattle and by those beta agonists that either increase protein synthesis or decrease catabolism.
|
Implications
|
|---|
It is generally assumed that feed efficiency decreases when gain is composed of fat rather than lean tissue. This study contradicts that assumption because it was observed that feed efficiency was the same over a range of 44 to 58% fat gain/empty body gain. The relative efficiency of converting NEg to fat compared with protein was 3.98; however, because the energy cost of lean and fat tissue gain seemed to be nearly equal, continued use of a single term, NEg, to calculate nutrient requirements seems correct. There was no correlation between ultrasound estimates of backfat thickness on the live animal and future feed efficiency. Many models assume that USDA Choice is attained at approximately 27% empty body fat. Even though carcass backfat thickness is highly correlated to carcass composition, its relationship with carcass marbling was poor (r2 = 0.08). Genetic selection for leaner carcasses with improved marbling score seems to have invalidated a relationship between percentage of empty body fat and marbling score.
1 Correspondence: 1232 240th Ave. (phone: 785-625-3425; fax: 785-623-4369; e-mail:Jbrethou{at}ksu.edu).
Received for publication September 26, 2003.
Accepted for publication July 21, 2004.
|
Literature Cited
|
|---|
Baldwin, R. L., N. E. Smith, J. Taylor, and M. Sharp. 1980. Manipulating metabolic parameters to improve growth rate and milk secretion. J. Anim. Sci. 51:1416–1428.
Brethour, J. R. 1997. Use of Ultrasonic Backfat Measures to Estimate Carcass Composition. Roundup 1997. Pages 10–14 in Kansas Agric. Exp. Stn. Rpt. of Prog. 784. Manhattan, KS.
Brethour, J. R. 2000. Using serial ultrasound measures to generate models of marbling and backfat thickness changes in feedlot cattle. J. Anim. Sci. 78:2055–2061.[Abstract/Free Full Text]
CSIRO. 1990. Feeding Standards for Australian Livestock. Ruminants. Commonwealth Scientific and Industrial Organization Research Organization. Melbourne, Australia.
Ferrell, C. L., and T. G. Jenkins. 1985. Energy utilization by Hereford and Simmental males and females. Anim. Prod. 41:53–61.
Geay, Y. 1984. Energy and protein utilization in growing cattle. J. Anim. Sci. 58:766–778.
Greiner, S. P., G. H. Rouse, D. E. Wilson, L. V. Cundiff, and T. E. Wheeler. 2003. Prediction of retail product weight and percentage using ultrasound and carcass measurements in beef cattle. J. Anim. Sci. 81:1736–1742.[Abstract/Free Full Text]
Guiroy, P. J., D. G. Fox, L. O. Tedeschi, M. J. Baker, and M. D. Cravey. 2001. Predicting individual feed requirements of cattle fed in groups. J. Anim. Sci. 79:1983–1995.[Abstract/Free Full Text]
Hollander, M., and D. A. Wolfe. 1973. Page 33 in Nonparametric Statistical Methods. John Wiley and Sons, New York.
Koontz, S. R., D. L. Hoag, J. L. Walker, and J. R. Brethour. 2000. Returns to market timing and sorting of fed cattle. In Proc. 2000 NCR-134 Conf. on Appl. Commod. Price Anal., Forecasting and Market Risk Manag., Chicago, IL.
Lobley, C. E. 1992. Control of the metabolic fate of amino acids: A review. J. Anim. Sci. 70:3264–3275.[Abstract]
Loveday, H. D., and M. E. Dikeman. 1980. Diet energy and adipose composition, lipogenesis and carcass composition. J. Anim. Sci. 51:78–88.[Abstract/Free Full Text]
NRC. 1996. Nutrient Requirements of Beef Cattle. 7th ed. Natl. Acad. Press, Washington, DC.
Rattray, P. V., and J. P. Joyce. 1976. Utilisation of metabolizible energy for fat and protein deposition in sheep. N. Z. J. Agric. Res. 19:299–305.
Reid, J. T., G. H. Wellington, and H. O. Dunn. 1955. Some relationships among the major chemical components of the bovine body and their application to nutritional investigations. J. Dairy Sci. 38:1344–1348.[Abstract/Free Full Text]
Snedecor, G. W. 1955. Page 346 in Statistical Methods. Iowa State College Press, Ames.
Tedeschi, L. O., D. G. Fox, and P. J. Guiroy. 2004. A decision support system to improve individual cattle management. 1. A mechanistic, dynamic model for animal growth. Agric. Syst. 79:171–204.
This article has been cited by other articles:
|
|
|
|
 
M. J. McPhee, J. W. Oltjen, T. R. Famula, and R. D. Sainz
Meta-analysis of factors affecting carcass characteristics of feedlot steers
J Anim Sci,
November 1, 2006;
84(11):
3143 - 3154.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. J. Baker, L. O. Tedeschi, D. G. Fox, W. R. Henning, and D. J. Ketchen
Using ultrasound measurements to predict body composition of yearling bulls
J Anim Sci,
October 1, 2006;
84(10):
2666 - 2672.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
