- Email: [email protected]
Prediction of true metabolizable energy from chemical composition of wheat milling by-products for ducks1
Metabolism and Nutrition Prediction of true metabolizable energy from chemical composition of wheat milling by-products for ducks1 H. F. Wan, W. Chen, Z. L. Qi, P. Peng, and J. Peng2 College of Animal Science and Technology, Huazhong Agricultural University, Wuhan, 430070 P. R. China
Key words: duck, true metabolizable energy, chemical composition, wheat milling by-product 2009 Poultry Science 88:92–97 doi:10.3382/ps.2008-00160 of aleurone and germ layer, leading to large variability in nutrient composition and ME of wheat by-products (Cromwell et al., 2000). Therefore, traditional diet formulating for ducks based on the tabulated TME values of wheat by-products is inadequate if an available formulation diet for practical production is to be obtained. Considering the considerable time and labor cost of in vivo determination of practical chemical composition of wheat by-products, there has been growing interest in the prediction of energy value of wheat by-products by establishing relationships with chemical analyses that are easily performed in the production. Most works report that fiber content shows the greatest but negative correlation with DE of diet in rabbit (Fernández-Carmona et al., 1996) and dog (Kienzle et al., 2006), giving rise to equations in which acid detergent fiber (ADF) or crude fiber (CF) are the main independent variables. This result strongly suggests that the fiber fraction should be an important consideration when the chemical composition of feedstuff is used to establish a regression equation for predicting the ME of feedstuff (Noblet and Perez., 1993; Fairbairn et al., 1999). Furthermore, Dale (1996) reported that
INTRODUCTION Wheat bran and shorts are the most common wheat milling by-products and widely used in commercial animal production in China. But there is a large variation in the chemical composition of wheat by-products because of varied wheat sources (e.g., soft vs. hard; Kim et al., 2005) and difference in processing techniques (e.g., fixed-system technique vs. setting-changed system; Li and Posner, 1989). It is reported that the starch, CP, and fiber among the proximate composition of wheat are the most variable components according to different varieties, growing locations, and climate, which in turn causes variability of ME value of wheat by-products (Kim et al., 2005). In addition, the difference in processing techniques causes a great range in the content ©2009 Poultry Science Association Inc. Received April 18, 2008. Accepted August 24, 2008. 1 Supported by the Key Scientific and Technological Project of Hubei Province in the Eleventh Five-Year, P. R. China (grant No. 2006AA202A04; 2006AA201B03). 2 Corresponding author: [email protected]
92
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
vs. 2.82%). The crude fiber, NDF, acid detergent fiber were highly but negatively correlated with TME (P < 0.01), in which the greatest correlation coefficient (r = −0.969) was observed between NDF and TME. No significant correlation of CP, EE, ash, and GE to TME was found among the 7 representative samples. The optimal equation in terms of R2 from using a single chemical analysis was obtained in the total group: TME = −0.1564NDF + 17.4696 (R2 = 0.94, P = 0.0003), and the TME prediction equation was improved by the addition of the EE and CP content to sequential analysis: TME = −0.17NDF + 0.98EE − 0.27CP + 19.31 (R2 = 0.99, residual SD = 0.35, P < 0.01). The results of present study suggest that NDF could be used as an effective indicator for the prediction of the TME value of wheat by-products for ducks.
ABSTRACT The present study was conducted to evaluate the effect of chemical composition of wheat by-products on the TME value to ducks and to establish the prediction models estimating TME. Seven representative samples were selected from 23 wheat by-products millings samples based on the neutral detergent fiber (NDF) content. According to the Sibbald method, male Cherry Valley ducks were chosen to assay the TME of 7 representative samples. Stepwise regression analysis was performed to establish the prediction equations of TME using CP, ether extract (EE), NDF, acid detergent fiber, crude fiber, and gross energy (GE) as independent variables. The NDF, CP, and DM of 23 samples of wheat by-product averaged to be 33.39 ± 11.04%, 19.2 ± 3.25%, 87.17 ± 0.95%, respectively. The TME values of 7 representative samples averaged 12.02 MJ/kg, with much larger CV than GE (17.72
PREDICTION OF TRUE METABOLIZABLE ENERGY
neutral detergent fiber (NDF) was the only component of proximate composition significantly correlated with TME of wheat by-products for broilers, indicating that NDF is probably a useful indicator for predicting TME of wheat by-products. The objective of present study were 1) to determine the proximate composition of 7 representative samples of 23 wheat by-products samples and 2) to establish relatively accurate prediction models estimating the TME values from chemical composition of wheat byproducts for ducks.
Sample Collection and Chemical Analysis Twenty-three samples of wheat milling by-products (including wheat bran and wheat shorts) were collected from 15 commercial firms located in the Henan, Anhui, and Hubei provinces in China. The samples were ground through a 1-mm screen and then stored for further analysis. Chemical analyses followed methods of the AOAC (2000) for DM (method 934.01) and CP (method 955.04); NDF was analyzed following Van Soest et al. (1991).
Representative Samples Selection and Chemical Analysis Seven feedstuff samples differing in NDF content were selected from 23 samples of wheat by-products to ensure that NDF content is within the range of means ± 2 SD (NDF ranging from 16.6 to 52.9%), details of sample selection are as follows: A, the smallest NDF value, B, C, and D, near the NDF mean; E and F, greater than the mean NDF values; and G, the greatest NDF value. According to the method described by AOAC (2000), chemical analyses were conducted for determination of DM (method 934.01), CP (method 955.04), ether extract (EE; method 920.39), ash (method 942.05), and CF (method 962.09) of samples selected. The gross energy (GE) of samples was determined by bomb calorimetry using Parr 1261 adiabatic calorimeter (Parr Instruments Co., Moline, IL). The NDF and ADF analyses were conducted following the procedure of Van Soest et al. (1991) and Van Soest (1973).
Experimental Design Seventy-two adult male meat ducks (Cherry Valley duck), with an average BW of 2.7 kg, were randomly assigned to 9 treatments of 8 replicates of 1 duck each. Each of the 7 wheat by-product samples was force-fed to each of 8 ducks, whereas corn starch was specifically provided to another 8 ducks; the remaining ducks were
fasted for determination of endogenous energy loss. Each duck was caged individually in the same room throughout the experiment.
Birds Feeding and Excreta Collection Force-feeding trials were carried out according to the modified method of Sirbald (1976). The protocol of this trial was approved by the Animal Care and Use Committee of College of Animal Science and Technology, Huazhong Agricultural University, and was carried out in accordance with the National Research Council’s Guide for the Care and Use of Laboratory Animals. Birds were fitted with individual excreta collection vessel as described previously by Ragland et al. (1997). In summary, approximately 48 h before the start of an experiment, each duck was surgically fitted with a bottle retainer lid. The feathers around the vent of ducks were plucked by hand. To avoid causing a laxative effect on ducks when using a single ingredient of the feeding diet, the D, E, F, and G feedstuff samples (NDF >32%) were mixed with corn starch at the ratio of 6 (feedstuff sample) to 4 (corn starch). After 36 h of fasting, 60 g of each feedstuff was force-fed to the birds through a 35-cm-long, 8-mm diameter tube. Total collection of excreta was initiated for 36 h after feeding. With regard to the determination of endogenous energy loss, 8 ducks were fasted for 72 h, and the excreta was collected for 36 h. Ten milliliters of 10% hydrochloride was added to the excreta of each duck during the excreta collecting period. Excreta samples from ducks that did not vomit were considered to be valid, of which at least 6 excreta samples for each tested feed sample were ensured. After being dried at 65°C and weighed, the excreta samples were ground through a 60-mesh screen and stored for GE and DM assays. The TME was calculated as follows: TME = [EI − (EO − EEL)]/FI, where EI equals GE intake of the feedstuff, EO equals GE voided of the feedstuff, EEL equals the endogenous energy loss; and FI equals feed intake of the feedstuff fed to each duck (60 g).
Statistical Analysis Statistical analysis was performed using SAS (SAS Institute, 1990). General linear model procedure was applied to study the effect of chemical composition on the TME. Correlation and sequential multiple linear regression analysis (stepwise procedure) were employed using DM, CP, CF, ADF, NDF, EE, GE, and ash as independent variables and TME as the dependent variable. The variance was considered to be significant when P < 0.05. In the proposed equations, the inclusion of independent variables was only considered when they caused a significant improvement (P < 0.05). The equations with the least residual SD (RSD) or those indicating the limits of some chemical criteria were presented.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
MATERIALS AND METHODS
93
94
Wan et al.
Table 1. The neutral detergent fiber (NDF), CP, and DM content of 23 wheat by-products samples (% on DM basis) Item
n
DM
CP
NDF
Mean Range among sources Minimum Maximum CV
23
87.17
19.20
33.39
85.16 89.86 1.09
14.64 27.03 16.92
16.60 52.91 33.06
rather than NDF alone, with the R2 increased up to 0.99, whereas the RSD decreased to 0.35.
DISCUSSION Variation of Chemical Compositions and TME Among Wheat By-Products
RESULTS
Table 2. Chemical composition of representative wheat by-products samples and TME value for ducks (% on DM basis)1 Item
NDF
ADF
CF
DM
CP
EE
Ash
GE (MJ/kg)
TME (MJ/kg)
A B C D E F G Mean SD CV
16.60 22.09 30.63 32.42 41.55 47.45 52.91 34.81 13.22 37.99
3.52 5.73 11.82 8.20 12.41 11.81 14.66 9.74 4.02 41.32
2.93 3.67 6.70 5.15 9.64 8.13 11.68 6.84 3.19 46.65
87 89 87 87 88 87 87 87 0.79 0.90
16.45 18.71 19.19 20.05 18.12 18.77 17.44 18.39 1.18 6.43
3.24 4.22 3.15 4.10 4.04 4.71 3.90 3.91 0.55 14.07
2.22 3.32 9.12 4.59 5.86 6.74 6.65 5.50 2.32 42.27
18.91 18.91 18.17 19.21 18.85 19.44 19.89 19.05 0.54 2.82
15.06 14.33 11.82 12.08 11.55 10.50 8.83 12.02 2.13 17.75
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
The chemical compositions of test ingredients (23 samples) were summarized in Table 1. The DM content was 87.7%, and the CP and NDF content on a DM basis was 19.2 and 33.39%, respectively. There was high CV in NDF (CV = 33.06%), suggesting a large range in the fiber content of wheat by-products. To establish accurate prediction equations of TME using fiber fractions as independent variable, 7 representative samples differing in the NDF content were selected for further chemical analysis and TME evaluation. Chemical composition values of wheat by-products were presented in Table 2; although little variation occurred in the CP, DM, and GE, a great range for some other values was observed, especially noticeable for CF (2.93 to 11.68%), NDF (16.60 to 52.91%), ADF (3.52 to 14.66%), ash (2.22 to 9.12%), and TME (8.83 to 15.06 MJ·kg−1). Accordingly, high CV was also detected for these values. The simple correlation matrices analyses were conducted to evaluate the correlation between chemical composition and TME values. As shown in Table 3, CF, NDF, and ADF among 7 representative wheat byproducts samples were highly but negatively correlated with TME, and the greatest correlation coefficient was observed between NDF and TME (r = −0.969; P < 0.001), whereas no significant correlation of CP, EE, and ash to TME was found among 7 representative wheat by-products samples. Equations obtained from multi-step regression analysis indicated that NDF was the best single predictor for TME (R2 = 0.948; Table 4). The accuracy of prediction for TME could be improved by adding EE and CP as predictor variables
In current study, a large range was observed in NDF content among chemical composition of 23 wheat byproducts samples from the different mills and places. Of the sample in Table 2, the greatest NDF was 3.19 times the least NDF. The numbers for ADF and CF are 4.16 and 3.99. However, the DM and CP content varied in a relatively narrow range, which is similar to that published by Cromwell et al. (2000) and Blas et al. (2000). Data from more detailed chemical analysis in 7 representative samples showed that, except for CP, DM, EE, there was a large range in the content of NDF, CF, ADF, indicating that the fiber fraction in the wheat by-products is more susceptible to varied sources of feedstuffs than other proximate components. Similar to NDF, a large range among samples was also observed in CF, reflected in the highest CV of CF. However, there was no significant correlation between CF and NDF, which is probably due to the discrepancy in the content of hemicellulose in wheat by-products of varied sources. In spite of low variability in GE, the TME among samples varied in a large range similar to NDF. As expected, the greater the fiber content, the less the AME values of wheat by-product. All the values, except for EE, show negative correlation with TME, in which the CF, NDF, ADF are highly correlated with TME (P < 0.001; Table 3). This result is in general agreement with other report showing that variable fiber fraction probably is the predominant factor in influencing the TME of wheat by-products for duck (Villamide and San Juan, 1998). Wheat milling by-products are produced when wheat is processed into flour, which is defined by AAFCO (2000) as fine particle of wheat bran, wheat shorts, wheat germ, and wheat flour. There are many factors in affecting the chemical composition and nutritional
95
PREDICTION OF TRUE METABOLIZABLE ENERGY Table 3. Correlation coefficients among representative wheat by-products samples in chemical composition and TME1 Correlation matrix CP EE Ash CF NDF ADF GE TME
CP
EE
Ash
CF
NDF
ADF
GE
TME
1.000
0.355
0.384 −0.049
−0.054 0.263 0.650
0.104 0.499 0.615 0.943**
0.185 0.226 0.842* 0.951*** 0.916**
−0.212 0.579 −0.159 0.471 0.619 0.290
−0.174 0.341 −0.722 −0.930** −0.969*** −0.9470** −0.543 1.000
value of wheat, including growing condition, wheat type, and postharvest storage (Gutiérrez-Alamo et al., 2008). The fiber fraction of wheat by-products probably had a wide range of variation because of the wheat variety (Kim et al., 2005), and the endosperm and bran content in wheat by-products varied considerably because of differences in processing technique (Huang et al., 1999). Wheat type and chemical composition are the most important factors affecting the TME value of wheat. For instance, soft wheat varieties tend to have greater starch and greater starch digestibility and AMEn for broiler than hard-wheat varieties (Gutiérrez-Alamo et al., 2008). However, the effect of wheat type on the TME value of wheat by-products should be evaluated in future research. With regard to the significant correlation of CF, NDF, and ADF to TME in current study, fiber is probably the most important factor affecting TME.
Establishment of Prediction Equations of TME from NDF To obtain equations to predict the TME of wheat byproduct, stepwise regression analyses were performed using a different chemical composition value. The best single predictor of TME was NDF, which explained 94% of TME variation (Table 4), suggesting that the NDF fraction among chemical compositions has the greatest single impact on influencing the TME content. Energy levels in barley have also been successfully estimated, with NDF as the primary predictor variable in a previous report (Morgan et al., 1987). Previous work con-
ducted on the wheat and wheat by-products shows that fiber content is the optimal predictor for predicting the ME and DE (Dale, 1996; Fairbairn et al., 1999; Blas et al., 2000; Noblet and Goff., 2001). In current study, we demonstrated that NDF among the fiber fraction could be used as an effective indicator affecting the TME of wheat by-products for ducks. In the method of NDF analysis proposed by Van Soest, most of soluble nutrient were dissolved in neutral detergent, and the neutral detergent residue (i.e., NDF fraction) comprises a heterogenous mixture of structural (cellulose and most of hemicellulose) and nonstructural polysaccharides, lignin, and a few of protein and materials connected with cell wall (Chesson and Austin, 1998; Fang et al., 2007a,b, 2008). It is commonly accepted that the NDF represent most of the fiber fraction of feedstuff, which are poorly digested by poultry like ducks because of the lack of enzymes that can effectively degrade these complex carbohydrates (Villamide and San Juan, 1998; Adeola, 2006; Fang et al., 2008). It is also supported by the report that the digestibility of fiber in grain is less than 20% in poultry (Terence et al., 2000). Therefore, energy utilization efficiency of wheat by-product is negatively and prominently affected by the NDF content, which serves to make NDF an effective predictor of TME of wheat byproduct for duck. For wheat by-products, NDF fraction could represent the most of indigestible fiber fraction because of very little pectin in Gramineae plant (Van Soest et al., 1991). This is, however, not the same case as in soy because the substantial quantities of pectin in legumes (Gidenne, 2003), which are known to decrease the ME of diet (Carré, 1993), were not included in the NDF fraction (Van Soest et al., 1991).
Table 4. Prediction of TME from chemical composition of wheat by-products1 Number in model
Regressive equation
R2
RSD
P
1 2 3
TME = −0.1564NDF + 17.4696 TME = −0.1718NDF + 0.7379EE + 15.1187 TME = −0.1742NDF + 0.9754EE − 0.2738CP + 19.3082
0.94 0.97 0.99
0.52 0.39 0.24
0.0003 0.0011 0.0025
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy. *P < 0.05; **P < 0.01; ***P < 0.001.
96
Wan et al.
REFERENCES AAFCO. 2000. Official Publication of Association of American Feed Control Officials, Oxford, IN. Adeola, O. 2006. Review of research in duck nutrient utilization. Int. J. Poult. Sci. 5:201–208. AOAC. 2000. Official Methods of Analysis. 17th ed. Assoc. Off. Anal. Chem., Arlington, VA. Blas, E., J. Fernandez-Carmona, C. Cervera, and J. J. Pascual. 2000. Nutritive value of coarse and fine wheat brans for rabbits. Anim. Feed Sci. Technol. 88:239–251.
Carré, B. 1993. Digestibility of carbohydrates in poultry. Pages 148– 163 in 9th Eur. Symp. Poult. Nutr. WPSA, Jelena, Poland. Chesson, A., and S. Austin. 1998. Defining the nutritional characteristics of plant carbohydrates—The non-ruminant perspective. Pages 205–210 in Proc. 19th Western Nutr. Conf., Saskatoon, Canada. Cromwell, G. L., T. R. Cline, J. D. Crenshaw, T. D. Crenshaw, R. A. Easter, R. C. C. R. Hamilton, G. M. Hill, A. J. Lewis, D. C. Mahan, J. L. Nelssen, J. E. Pettigrew, T. L. Veum, and J. T. Yen. 2000. Variability among sources and laboratories in analyses of wheat middlings. J. Anim. Sci. 78:2652–2658. Dale, N. 1996. The metabolizable energy of wheat by-products. J. Appl. Poult. Res. 5:105–108. Fairbairn, S. L., J. F. Patience, H. L. Classen, and R. T. Zijlstra.1999. The energy content of barley fed to growing pigs: Characterizing the nature of its variability and developing prediction equations for its estimation. J. Anim. Sci. 77:1502–1512. Fang, Z. F., Z. L. Liu, H. Y. Qian, Z. L. Qi, L. B. Ma, and J. Peng. 2008. Effects of enzyme addition on the nutritive value of broiler diets containing hulled or dehulled Chinese double-low rapeseed meals. J. Anim. Physiol. Anim. Nutr. (Berl.) doi:10.1111/j.14390396.2008.00829.x. Fang, Z. F., J. Peng, Z. L. Liu, and Y. G. Liu. 2007a. Responses of non-starch polysaccharide-degrading enzymes on digestibility and performance of growing pigs fed a diet based on corn, soybean meal and Chinese double-low rapeseed. J. Anim. Physiol. Anim. Nutr. (Berl.) 91:361–368. Fang, Z. F., J. Peng, T. J. Tang, Z. L. Liu, and J. J. Dai. 2007b. Xylanase supplementation improved digestibility and performance of growing pigs fed Chinese double-low rapeseed meal inclusion diets: In vitro and in vivo studies. Asian-australas. J. Anim. Sci. 20:1721–1728. Fernández-Carmona, J., C. Cervera, and E. Blas. 1996. Prediction of the energy value of rabbit feeds varying widely in fibre content. Anim. Feed Sci. Technol. 64:61–75. Gidenne, T. 2003. Fibres in rabbit feeding for digestive troubles prevention: Respective role of low-digested and digestible fibre. Livest. Prod. Sci. 81:105–117. Gutiérrez-Alamo, A., P. Pérez de ayala, M. W. A. Verstegen, L. A. Den hartog, and M. J. Villamide. 2008. Variability in wheat: Factors affecting its nutritional value. World’s Poult. Sci. J. 64:20–40. Huang, S. X., W. C. Sauer, and B. Marty. 2001. Ileal digestibilities of neutral detergent fiber, crude protein, and amino acids associated with neutral detergent fiber in wheat shorts for growing pigs. J. Anim. Sci. 79:2388–2396. Huang, S. X., W. C. Sauer, B. Marty, and R. T. Hardin. 1999. Amino acid digestibilities in different samples of wheat shorts for growing pigs. J. Anim. Sci. 77:2469–2477. Kienzle, E., B. Vincent, and S. Alexandra. 2006. Prediction of energy digestibility in complete dry foods for dogs and cats by total dietary fiber. J. Nutr. 136(Suppl.):2041–2044. Kim, J. C., P. H. Simmins, B. P. Mullan, and J. R. Pluske. 2005. The digestible energy value of wheat for pigs with special reference to the post-weaned animal. Anim. Feed Sci. Technol. 122:257–287. Li, Y. Z., and E. S. Posner. 1989. An experimental milling technique for various flour extraction levels. Cereal Chem. 66:324–328. Licitra, G., T. M. Hernandez, and P. J. Van Soest. 1996. Standardization of procedures for nitrogen fractionation of ruminants feeds. Anim. Feed Sci. Technol. 57:347–358. Morgan, C. A., C. T. Whittemore, P. Phillips, and P. Crooks. 1987. The prediction of the energy value of compounded pig foods from chemical analysis. Anim. Feed Sci. Technol. 17:81–107. Noblet, J., and L. G. Goff. 2001. Effect of dietary fibre on the energy value of feeds for pigs. Anim. Feed Sci. Technol. 90:35–52. Noblet, J., and J. M. Perez. 1993. Prediction of digestibility of nutrients and energy values of pig diets from chemical analysis. J. Anim. Sci. 71:3389–3398. Ragland, D., D. King, and O. Adeola. 1997. Determination of metabolizable energy contents of feed ingredients for ducks. Poult. Sci. 76:1287–1291. SAS Institute. 1990. User’s Guide – Version 6 Edition. SAS Institute Inc., Cary, NC.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
In current study, neither CF nor ADF, with a high degree of variability (CV = 46.65 and 41.32%, respectively), becomes the best single predictor of TME in the 1-dimensional linear regression equation. This is probably because of the CF and ADF characteristics. After being dealt with different detergents, most of the hemicelluloses and CP associated with NDF, which could not be further digested, are dissolved in the detergent solutions and removed in the CF and ADF analysis procedure (Van Soest et al., 1991). As such, ADF fraction comprises cellulose, lignin, acid detergent insoluble N, acid-insoluble ash, and silica. Accordingly, CF mainly contains cellulose and lignin. Both of these are not valid fiber fractions for nutritional use or for the prediction of TME for ducks because they underestimate the indigestible fiber fraction in the wheat by-products. As such, neither ADF nor CF could be the optimal predictor of TME of wheat by-product for ducks. A significant improvement in the accuracy of prediction was obtained in equations 2 and 3 (Table 4) when EE and CP were introduced into the prediction model, increasing R2 from 0.94 up to 0.99, and decreasing RSD from 0.52 to 0.24. The EE was positively correlated to TME although the correlation coefficient was less than that of some fiber measurements (r = 0.3411 vs. 0.9296, 0.9694, and 0.9469 for CF, NDF, and ADF, respectively). It is suggested that EE, as an important energy source, displays a positive contribution to TME of wheat by-products. In contrast, CP is negatively correlated with TME, which can be partly explained by the report that CP associated with neutral detergent fiber that represents indigestible CP negatively affect the TME of wheat by-product for duck (Licitra et al., 1996; Huang et al., 2001). In conclusion, wheat middlings is one of the feed ingredients that is extremely variable in proximate components, particularly the fiber content. The NDF among the chemical compositions is the key factor in the selection of a better prediction equation of TME for wheat by-products. Also, the accuracy of prediction equation of TME could be further improved when other chemical characteristics are taken into consideration. The results of this study suggest that it is possible to develop an accurate equation to predict the TME of feedstuffs from chemical compositions. The establishment of prediction equation of TME would, to some extent, improve the accuracy and practicability of diet formulation in commercial animal production.
PREDICTION OF TRUE METABOLIZABLE ENERGY Sirbald, I. R. 1976. A bioassay for true metabolizable energy in feedingstuffs. Poult. Sci. 55:303–308. Terence, J. D., J. W. Peter, M. Adam, F. D. Fanning, and R. D. William. 2000. Digestive function in Australian magpie geese (Anseranas semipalmata). Aust. J. Zool. 48:265–279. Van Soest, P. J. 1973. Collaborative study of acid detergent fiber and lignin. J. Assoc. Off. Anal. Chem. 56:781–784.
97
Van Soest, P. J., J. B. Robertson, and B. A. Lewis. 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74:3583– 3597. Villamide, M. J., and L. D. San Juan. 1998. Effect of chemical composition of sunflower seed meal on its true metabolizable energy and amino acid digestibility. Poult. Sci. 77:1884–1892.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
Key words: duck, true metabolizable energy, chemical composition, wheat milling by-product 2009 Poultry Science 88:92–97 doi:10.3382/ps.2008-00160 of aleurone and germ layer, leading to large variability in nutrient composition and ME of wheat by-products (Cromwell et al., 2000). Therefore, traditional diet formulating for ducks based on the tabulated TME values of wheat by-products is inadequate if an available formulation diet for practical production is to be obtained. Considering the considerable time and labor cost of in vivo determination of practical chemical composition of wheat by-products, there has been growing interest in the prediction of energy value of wheat by-products by establishing relationships with chemical analyses that are easily performed in the production. Most works report that fiber content shows the greatest but negative correlation with DE of diet in rabbit (Fernández-Carmona et al., 1996) and dog (Kienzle et al., 2006), giving rise to equations in which acid detergent fiber (ADF) or crude fiber (CF) are the main independent variables. This result strongly suggests that the fiber fraction should be an important consideration when the chemical composition of feedstuff is used to establish a regression equation for predicting the ME of feedstuff (Noblet and Perez., 1993; Fairbairn et al., 1999). Furthermore, Dale (1996) reported that
INTRODUCTION Wheat bran and shorts are the most common wheat milling by-products and widely used in commercial animal production in China. But there is a large variation in the chemical composition of wheat by-products because of varied wheat sources (e.g., soft vs. hard; Kim et al., 2005) and difference in processing techniques (e.g., fixed-system technique vs. setting-changed system; Li and Posner, 1989). It is reported that the starch, CP, and fiber among the proximate composition of wheat are the most variable components according to different varieties, growing locations, and climate, which in turn causes variability of ME value of wheat by-products (Kim et al., 2005). In addition, the difference in processing techniques causes a great range in the content ©2009 Poultry Science Association Inc. Received April 18, 2008. Accepted August 24, 2008. 1 Supported by the Key Scientific and Technological Project of Hubei Province in the Eleventh Five-Year, P. R. China (grant No. 2006AA202A04; 2006AA201B03). 2 Corresponding author: [email protected]
92
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
vs. 2.82%). The crude fiber, NDF, acid detergent fiber were highly but negatively correlated with TME (P < 0.01), in which the greatest correlation coefficient (r = −0.969) was observed between NDF and TME. No significant correlation of CP, EE, ash, and GE to TME was found among the 7 representative samples. The optimal equation in terms of R2 from using a single chemical analysis was obtained in the total group: TME = −0.1564NDF + 17.4696 (R2 = 0.94, P = 0.0003), and the TME prediction equation was improved by the addition of the EE and CP content to sequential analysis: TME = −0.17NDF + 0.98EE − 0.27CP + 19.31 (R2 = 0.99, residual SD = 0.35, P < 0.01). The results of present study suggest that NDF could be used as an effective indicator for the prediction of the TME value of wheat by-products for ducks.
ABSTRACT The present study was conducted to evaluate the effect of chemical composition of wheat by-products on the TME value to ducks and to establish the prediction models estimating TME. Seven representative samples were selected from 23 wheat by-products millings samples based on the neutral detergent fiber (NDF) content. According to the Sibbald method, male Cherry Valley ducks were chosen to assay the TME of 7 representative samples. Stepwise regression analysis was performed to establish the prediction equations of TME using CP, ether extract (EE), NDF, acid detergent fiber, crude fiber, and gross energy (GE) as independent variables. The NDF, CP, and DM of 23 samples of wheat by-product averaged to be 33.39 ± 11.04%, 19.2 ± 3.25%, 87.17 ± 0.95%, respectively. The TME values of 7 representative samples averaged 12.02 MJ/kg, with much larger CV than GE (17.72
PREDICTION OF TRUE METABOLIZABLE ENERGY
neutral detergent fiber (NDF) was the only component of proximate composition significantly correlated with TME of wheat by-products for broilers, indicating that NDF is probably a useful indicator for predicting TME of wheat by-products. The objective of present study were 1) to determine the proximate composition of 7 representative samples of 23 wheat by-products samples and 2) to establish relatively accurate prediction models estimating the TME values from chemical composition of wheat byproducts for ducks.
Sample Collection and Chemical Analysis Twenty-three samples of wheat milling by-products (including wheat bran and wheat shorts) were collected from 15 commercial firms located in the Henan, Anhui, and Hubei provinces in China. The samples were ground through a 1-mm screen and then stored for further analysis. Chemical analyses followed methods of the AOAC (2000) for DM (method 934.01) and CP (method 955.04); NDF was analyzed following Van Soest et al. (1991).
Representative Samples Selection and Chemical Analysis Seven feedstuff samples differing in NDF content were selected from 23 samples of wheat by-products to ensure that NDF content is within the range of means ± 2 SD (NDF ranging from 16.6 to 52.9%), details of sample selection are as follows: A, the smallest NDF value, B, C, and D, near the NDF mean; E and F, greater than the mean NDF values; and G, the greatest NDF value. According to the method described by AOAC (2000), chemical analyses were conducted for determination of DM (method 934.01), CP (method 955.04), ether extract (EE; method 920.39), ash (method 942.05), and CF (method 962.09) of samples selected. The gross energy (GE) of samples was determined by bomb calorimetry using Parr 1261 adiabatic calorimeter (Parr Instruments Co., Moline, IL). The NDF and ADF analyses were conducted following the procedure of Van Soest et al. (1991) and Van Soest (1973).
Experimental Design Seventy-two adult male meat ducks (Cherry Valley duck), with an average BW of 2.7 kg, were randomly assigned to 9 treatments of 8 replicates of 1 duck each. Each of the 7 wheat by-product samples was force-fed to each of 8 ducks, whereas corn starch was specifically provided to another 8 ducks; the remaining ducks were
fasted for determination of endogenous energy loss. Each duck was caged individually in the same room throughout the experiment.
Birds Feeding and Excreta Collection Force-feeding trials were carried out according to the modified method of Sirbald (1976). The protocol of this trial was approved by the Animal Care and Use Committee of College of Animal Science and Technology, Huazhong Agricultural University, and was carried out in accordance with the National Research Council’s Guide for the Care and Use of Laboratory Animals. Birds were fitted with individual excreta collection vessel as described previously by Ragland et al. (1997). In summary, approximately 48 h before the start of an experiment, each duck was surgically fitted with a bottle retainer lid. The feathers around the vent of ducks were plucked by hand. To avoid causing a laxative effect on ducks when using a single ingredient of the feeding diet, the D, E, F, and G feedstuff samples (NDF >32%) were mixed with corn starch at the ratio of 6 (feedstuff sample) to 4 (corn starch). After 36 h of fasting, 60 g of each feedstuff was force-fed to the birds through a 35-cm-long, 8-mm diameter tube. Total collection of excreta was initiated for 36 h after feeding. With regard to the determination of endogenous energy loss, 8 ducks were fasted for 72 h, and the excreta was collected for 36 h. Ten milliliters of 10% hydrochloride was added to the excreta of each duck during the excreta collecting period. Excreta samples from ducks that did not vomit were considered to be valid, of which at least 6 excreta samples for each tested feed sample were ensured. After being dried at 65°C and weighed, the excreta samples were ground through a 60-mesh screen and stored for GE and DM assays. The TME was calculated as follows: TME = [EI − (EO − EEL)]/FI, where EI equals GE intake of the feedstuff, EO equals GE voided of the feedstuff, EEL equals the endogenous energy loss; and FI equals feed intake of the feedstuff fed to each duck (60 g).
Statistical Analysis Statistical analysis was performed using SAS (SAS Institute, 1990). General linear model procedure was applied to study the effect of chemical composition on the TME. Correlation and sequential multiple linear regression analysis (stepwise procedure) were employed using DM, CP, CF, ADF, NDF, EE, GE, and ash as independent variables and TME as the dependent variable. The variance was considered to be significant when P < 0.05. In the proposed equations, the inclusion of independent variables was only considered when they caused a significant improvement (P < 0.05). The equations with the least residual SD (RSD) or those indicating the limits of some chemical criteria were presented.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
MATERIALS AND METHODS
93
94
Wan et al.
Table 1. The neutral detergent fiber (NDF), CP, and DM content of 23 wheat by-products samples (% on DM basis) Item
n
DM
CP
NDF
Mean Range among sources Minimum Maximum CV
23
87.17
19.20
33.39
85.16 89.86 1.09
14.64 27.03 16.92
16.60 52.91 33.06
rather than NDF alone, with the R2 increased up to 0.99, whereas the RSD decreased to 0.35.
DISCUSSION Variation of Chemical Compositions and TME Among Wheat By-Products
RESULTS
Table 2. Chemical composition of representative wheat by-products samples and TME value for ducks (% on DM basis)1 Item
NDF
ADF
CF
DM
CP
EE
Ash
GE (MJ/kg)
TME (MJ/kg)
A B C D E F G Mean SD CV
16.60 22.09 30.63 32.42 41.55 47.45 52.91 34.81 13.22 37.99
3.52 5.73 11.82 8.20 12.41 11.81 14.66 9.74 4.02 41.32
2.93 3.67 6.70 5.15 9.64 8.13 11.68 6.84 3.19 46.65
87 89 87 87 88 87 87 87 0.79 0.90
16.45 18.71 19.19 20.05 18.12 18.77 17.44 18.39 1.18 6.43
3.24 4.22 3.15 4.10 4.04 4.71 3.90 3.91 0.55 14.07
2.22 3.32 9.12 4.59 5.86 6.74 6.65 5.50 2.32 42.27
18.91 18.91 18.17 19.21 18.85 19.44 19.89 19.05 0.54 2.82
15.06 14.33 11.82 12.08 11.55 10.50 8.83 12.02 2.13 17.75
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
The chemical compositions of test ingredients (23 samples) were summarized in Table 1. The DM content was 87.7%, and the CP and NDF content on a DM basis was 19.2 and 33.39%, respectively. There was high CV in NDF (CV = 33.06%), suggesting a large range in the fiber content of wheat by-products. To establish accurate prediction equations of TME using fiber fractions as independent variable, 7 representative samples differing in the NDF content were selected for further chemical analysis and TME evaluation. Chemical composition values of wheat by-products were presented in Table 2; although little variation occurred in the CP, DM, and GE, a great range for some other values was observed, especially noticeable for CF (2.93 to 11.68%), NDF (16.60 to 52.91%), ADF (3.52 to 14.66%), ash (2.22 to 9.12%), and TME (8.83 to 15.06 MJ·kg−1). Accordingly, high CV was also detected for these values. The simple correlation matrices analyses were conducted to evaluate the correlation between chemical composition and TME values. As shown in Table 3, CF, NDF, and ADF among 7 representative wheat byproducts samples were highly but negatively correlated with TME, and the greatest correlation coefficient was observed between NDF and TME (r = −0.969; P < 0.001), whereas no significant correlation of CP, EE, and ash to TME was found among 7 representative wheat by-products samples. Equations obtained from multi-step regression analysis indicated that NDF was the best single predictor for TME (R2 = 0.948; Table 4). The accuracy of prediction for TME could be improved by adding EE and CP as predictor variables
In current study, a large range was observed in NDF content among chemical composition of 23 wheat byproducts samples from the different mills and places. Of the sample in Table 2, the greatest NDF was 3.19 times the least NDF. The numbers for ADF and CF are 4.16 and 3.99. However, the DM and CP content varied in a relatively narrow range, which is similar to that published by Cromwell et al. (2000) and Blas et al. (2000). Data from more detailed chemical analysis in 7 representative samples showed that, except for CP, DM, EE, there was a large range in the content of NDF, CF, ADF, indicating that the fiber fraction in the wheat by-products is more susceptible to varied sources of feedstuffs than other proximate components. Similar to NDF, a large range among samples was also observed in CF, reflected in the highest CV of CF. However, there was no significant correlation between CF and NDF, which is probably due to the discrepancy in the content of hemicellulose in wheat by-products of varied sources. In spite of low variability in GE, the TME among samples varied in a large range similar to NDF. As expected, the greater the fiber content, the less the AME values of wheat by-product. All the values, except for EE, show negative correlation with TME, in which the CF, NDF, ADF are highly correlated with TME (P < 0.001; Table 3). This result is in general agreement with other report showing that variable fiber fraction probably is the predominant factor in influencing the TME of wheat by-products for duck (Villamide and San Juan, 1998). Wheat milling by-products are produced when wheat is processed into flour, which is defined by AAFCO (2000) as fine particle of wheat bran, wheat shorts, wheat germ, and wheat flour. There are many factors in affecting the chemical composition and nutritional
95
PREDICTION OF TRUE METABOLIZABLE ENERGY Table 3. Correlation coefficients among representative wheat by-products samples in chemical composition and TME1 Correlation matrix CP EE Ash CF NDF ADF GE TME
CP
EE
Ash
CF
NDF
ADF
GE
TME
1.000
0.355
0.384 −0.049
−0.054 0.263 0.650
0.104 0.499 0.615 0.943**
0.185 0.226 0.842* 0.951*** 0.916**
−0.212 0.579 −0.159 0.471 0.619 0.290
−0.174 0.341 −0.722 −0.930** −0.969*** −0.9470** −0.543 1.000
value of wheat, including growing condition, wheat type, and postharvest storage (Gutiérrez-Alamo et al., 2008). The fiber fraction of wheat by-products probably had a wide range of variation because of the wheat variety (Kim et al., 2005), and the endosperm and bran content in wheat by-products varied considerably because of differences in processing technique (Huang et al., 1999). Wheat type and chemical composition are the most important factors affecting the TME value of wheat. For instance, soft wheat varieties tend to have greater starch and greater starch digestibility and AMEn for broiler than hard-wheat varieties (Gutiérrez-Alamo et al., 2008). However, the effect of wheat type on the TME value of wheat by-products should be evaluated in future research. With regard to the significant correlation of CF, NDF, and ADF to TME in current study, fiber is probably the most important factor affecting TME.
Establishment of Prediction Equations of TME from NDF To obtain equations to predict the TME of wheat byproduct, stepwise regression analyses were performed using a different chemical composition value. The best single predictor of TME was NDF, which explained 94% of TME variation (Table 4), suggesting that the NDF fraction among chemical compositions has the greatest single impact on influencing the TME content. Energy levels in barley have also been successfully estimated, with NDF as the primary predictor variable in a previous report (Morgan et al., 1987). Previous work con-
ducted on the wheat and wheat by-products shows that fiber content is the optimal predictor for predicting the ME and DE (Dale, 1996; Fairbairn et al., 1999; Blas et al., 2000; Noblet and Goff., 2001). In current study, we demonstrated that NDF among the fiber fraction could be used as an effective indicator affecting the TME of wheat by-products for ducks. In the method of NDF analysis proposed by Van Soest, most of soluble nutrient were dissolved in neutral detergent, and the neutral detergent residue (i.e., NDF fraction) comprises a heterogenous mixture of structural (cellulose and most of hemicellulose) and nonstructural polysaccharides, lignin, and a few of protein and materials connected with cell wall (Chesson and Austin, 1998; Fang et al., 2007a,b, 2008). It is commonly accepted that the NDF represent most of the fiber fraction of feedstuff, which are poorly digested by poultry like ducks because of the lack of enzymes that can effectively degrade these complex carbohydrates (Villamide and San Juan, 1998; Adeola, 2006; Fang et al., 2008). It is also supported by the report that the digestibility of fiber in grain is less than 20% in poultry (Terence et al., 2000). Therefore, energy utilization efficiency of wheat by-product is negatively and prominently affected by the NDF content, which serves to make NDF an effective predictor of TME of wheat byproduct for duck. For wheat by-products, NDF fraction could represent the most of indigestible fiber fraction because of very little pectin in Gramineae plant (Van Soest et al., 1991). This is, however, not the same case as in soy because the substantial quantities of pectin in legumes (Gidenne, 2003), which are known to decrease the ME of diet (Carré, 1993), were not included in the NDF fraction (Van Soest et al., 1991).
Table 4. Prediction of TME from chemical composition of wheat by-products1 Number in model
Regressive equation
R2
RSD
P
1 2 3
TME = −0.1564NDF + 17.4696 TME = −0.1718NDF + 0.7379EE + 15.1187 TME = −0.1742NDF + 0.9754EE − 0.2738CP + 19.3082
0.94 0.97 0.99
0.52 0.39 0.24
0.0003 0.0011 0.0025
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
1 NDF = neutral detergent fiber; ADF = acid detergent fiber; CF = crude fiber; EE = ether extract; GE = gross energy. *P < 0.05; **P < 0.01; ***P < 0.001.
96
Wan et al.
REFERENCES AAFCO. 2000. Official Publication of Association of American Feed Control Officials, Oxford, IN. Adeola, O. 2006. Review of research in duck nutrient utilization. Int. J. Poult. Sci. 5:201–208. AOAC. 2000. Official Methods of Analysis. 17th ed. Assoc. Off. Anal. Chem., Arlington, VA. Blas, E., J. Fernandez-Carmona, C. Cervera, and J. J. Pascual. 2000. Nutritive value of coarse and fine wheat brans for rabbits. Anim. Feed Sci. Technol. 88:239–251.
Carré, B. 1993. Digestibility of carbohydrates in poultry. Pages 148– 163 in 9th Eur. Symp. Poult. Nutr. WPSA, Jelena, Poland. Chesson, A., and S. Austin. 1998. Defining the nutritional characteristics of plant carbohydrates—The non-ruminant perspective. Pages 205–210 in Proc. 19th Western Nutr. Conf., Saskatoon, Canada. Cromwell, G. L., T. R. Cline, J. D. Crenshaw, T. D. Crenshaw, R. A. Easter, R. C. C. R. Hamilton, G. M. Hill, A. J. Lewis, D. C. Mahan, J. L. Nelssen, J. E. Pettigrew, T. L. Veum, and J. T. Yen. 2000. Variability among sources and laboratories in analyses of wheat middlings. J. Anim. Sci. 78:2652–2658. Dale, N. 1996. The metabolizable energy of wheat by-products. J. Appl. Poult. Res. 5:105–108. Fairbairn, S. L., J. F. Patience, H. L. Classen, and R. T. Zijlstra.1999. The energy content of barley fed to growing pigs: Characterizing the nature of its variability and developing prediction equations for its estimation. J. Anim. Sci. 77:1502–1512. Fang, Z. F., Z. L. Liu, H. Y. Qian, Z. L. Qi, L. B. Ma, and J. Peng. 2008. Effects of enzyme addition on the nutritive value of broiler diets containing hulled or dehulled Chinese double-low rapeseed meals. J. Anim. Physiol. Anim. Nutr. (Berl.) doi:10.1111/j.14390396.2008.00829.x. Fang, Z. F., J. Peng, Z. L. Liu, and Y. G. Liu. 2007a. Responses of non-starch polysaccharide-degrading enzymes on digestibility and performance of growing pigs fed a diet based on corn, soybean meal and Chinese double-low rapeseed. J. Anim. Physiol. Anim. Nutr. (Berl.) 91:361–368. Fang, Z. F., J. Peng, T. J. Tang, Z. L. Liu, and J. J. Dai. 2007b. Xylanase supplementation improved digestibility and performance of growing pigs fed Chinese double-low rapeseed meal inclusion diets: In vitro and in vivo studies. Asian-australas. J. Anim. Sci. 20:1721–1728. Fernández-Carmona, J., C. Cervera, and E. Blas. 1996. Prediction of the energy value of rabbit feeds varying widely in fibre content. Anim. Feed Sci. Technol. 64:61–75. Gidenne, T. 2003. Fibres in rabbit feeding for digestive troubles prevention: Respective role of low-digested and digestible fibre. Livest. Prod. Sci. 81:105–117. Gutiérrez-Alamo, A., P. Pérez de ayala, M. W. A. Verstegen, L. A. Den hartog, and M. J. Villamide. 2008. Variability in wheat: Factors affecting its nutritional value. World’s Poult. Sci. J. 64:20–40. Huang, S. X., W. C. Sauer, and B. Marty. 2001. Ileal digestibilities of neutral detergent fiber, crude protein, and amino acids associated with neutral detergent fiber in wheat shorts for growing pigs. J. Anim. Sci. 79:2388–2396. Huang, S. X., W. C. Sauer, B. Marty, and R. T. Hardin. 1999. Amino acid digestibilities in different samples of wheat shorts for growing pigs. J. Anim. Sci. 77:2469–2477. Kienzle, E., B. Vincent, and S. Alexandra. 2006. Prediction of energy digestibility in complete dry foods for dogs and cats by total dietary fiber. J. Nutr. 136(Suppl.):2041–2044. Kim, J. C., P. H. Simmins, B. P. Mullan, and J. R. Pluske. 2005. The digestible energy value of wheat for pigs with special reference to the post-weaned animal. Anim. Feed Sci. Technol. 122:257–287. Li, Y. Z., and E. S. Posner. 1989. An experimental milling technique for various flour extraction levels. Cereal Chem. 66:324–328. Licitra, G., T. M. Hernandez, and P. J. Van Soest. 1996. Standardization of procedures for nitrogen fractionation of ruminants feeds. Anim. Feed Sci. Technol. 57:347–358. Morgan, C. A., C. T. Whittemore, P. Phillips, and P. Crooks. 1987. The prediction of the energy value of compounded pig foods from chemical analysis. Anim. Feed Sci. Technol. 17:81–107. Noblet, J., and L. G. Goff. 2001. Effect of dietary fibre on the energy value of feeds for pigs. Anim. Feed Sci. Technol. 90:35–52. Noblet, J., and J. M. Perez. 1993. Prediction of digestibility of nutrients and energy values of pig diets from chemical analysis. J. Anim. Sci. 71:3389–3398. Ragland, D., D. King, and O. Adeola. 1997. Determination of metabolizable energy contents of feed ingredients for ducks. Poult. Sci. 76:1287–1291. SAS Institute. 1990. User’s Guide – Version 6 Edition. SAS Institute Inc., Cary, NC.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015
In current study, neither CF nor ADF, with a high degree of variability (CV = 46.65 and 41.32%, respectively), becomes the best single predictor of TME in the 1-dimensional linear regression equation. This is probably because of the CF and ADF characteristics. After being dealt with different detergents, most of the hemicelluloses and CP associated with NDF, which could not be further digested, are dissolved in the detergent solutions and removed in the CF and ADF analysis procedure (Van Soest et al., 1991). As such, ADF fraction comprises cellulose, lignin, acid detergent insoluble N, acid-insoluble ash, and silica. Accordingly, CF mainly contains cellulose and lignin. Both of these are not valid fiber fractions for nutritional use or for the prediction of TME for ducks because they underestimate the indigestible fiber fraction in the wheat by-products. As such, neither ADF nor CF could be the optimal predictor of TME of wheat by-product for ducks. A significant improvement in the accuracy of prediction was obtained in equations 2 and 3 (Table 4) when EE and CP were introduced into the prediction model, increasing R2 from 0.94 up to 0.99, and decreasing RSD from 0.52 to 0.24. The EE was positively correlated to TME although the correlation coefficient was less than that of some fiber measurements (r = 0.3411 vs. 0.9296, 0.9694, and 0.9469 for CF, NDF, and ADF, respectively). It is suggested that EE, as an important energy source, displays a positive contribution to TME of wheat by-products. In contrast, CP is negatively correlated with TME, which can be partly explained by the report that CP associated with neutral detergent fiber that represents indigestible CP negatively affect the TME of wheat by-product for duck (Licitra et al., 1996; Huang et al., 2001). In conclusion, wheat middlings is one of the feed ingredients that is extremely variable in proximate components, particularly the fiber content. The NDF among the chemical compositions is the key factor in the selection of a better prediction equation of TME for wheat by-products. Also, the accuracy of prediction equation of TME could be further improved when other chemical characteristics are taken into consideration. The results of this study suggest that it is possible to develop an accurate equation to predict the TME of feedstuffs from chemical compositions. The establishment of prediction equation of TME would, to some extent, improve the accuracy and practicability of diet formulation in commercial animal production.
PREDICTION OF TRUE METABOLIZABLE ENERGY Sirbald, I. R. 1976. A bioassay for true metabolizable energy in feedingstuffs. Poult. Sci. 55:303–308. Terence, J. D., J. W. Peter, M. Adam, F. D. Fanning, and R. D. William. 2000. Digestive function in Australian magpie geese (Anseranas semipalmata). Aust. J. Zool. 48:265–279. Van Soest, P. J. 1973. Collaborative study of acid detergent fiber and lignin. J. Assoc. Off. Anal. Chem. 56:781–784.
97
Van Soest, P. J., J. B. Robertson, and B. A. Lewis. 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74:3583– 3597. Villamide, M. J., and L. D. San Juan. 1998. Effect of chemical composition of sunflower seed meal on its true metabolizable energy and amino acid digestibility. Poult. Sci. 77:1884–1892.
Downloaded from http://ps.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on March 11, 2015