Distillers Dried Grains with Solubles (DDGS), a byproduct of ethanol production from corn that is becoming available in increasing quantities, is a potentially valuable ingredient for use in diets for swine. However, DDGS is very variable in composition both within and between sources (ethanol plants). This variability in nutritional value of DDGS reflects both variation in the composition of the corn going into the ethanol fermentation process and also variation in processing conditions. Economical use of any feed ingredient for pigs depends on accurate information being available on its nutrient composition and metabolizable energy (ME) content. In practice, the ME content of DDGS is generally estimated using prediction equations based on its chemical composition; however, published equations currently widely used by the industry give a range of ME values for the same sample of DDGS.

The objective of this study was to develop regression equations to predict the metabolizable energy (ME) content of DDGS based on chemical composition. The study used DDGS samples obtained from 17 sources (Midwestern ethanol plants) that were chosen to represent the variation in nutrient content currently available to the industry. The DDGS samples varied widely in particle size and those above a target particle size of 340 μm (15 of the 17 samples) were ground through a hammermill to a common particle size.

The DE and ME contents of the 18 experimental diets (a corn-based control and the 17 DDGS sample diets) were measured using a standard energy balance study involving 36 barrows housed individually in metabolism crates. The experimental diets were fed for a 7-day period which consisted of a 4-day adaptation period followed by a 3-day collection period during which total but separate collection of feces and urine was carried out. Gross energy of diets, feces and urine were determined by bomb calorimetry. Chemical composition (crude protein, crude fat, crude fiber, ADF, NDF, ash, and starch) of each DDGS sample was analyzed by two independent commercial laboratories. Equations to predict the ME content based on chemical composition and particle size after grinding were developed. To test for differences between laboratories, equations were developed based on the chemical analysis of each laboratory separately and for the average of the results of the two laboratories.

There was considerable variation in the energy content and chemical composition of the 17 DDGS samples. In addition, there were relatively large differences between the results of the chemical analysis for the two laboratories for a number of the chemical components, particularly, crude fat, ADF, NDF, and starch. The DE and ME values for the corn sample determined in the energy balance study (3,893 ± 71.4 and 3,813 ± 60.6 kcal ME/kg DM, respectively) were within the range of previously reported values for corn. The DE content of the DDGS samples, determined by the difference method, ranged from 3,663 to 4,107 (mean 3,954 ± 112.5) kcal/kg DM and the ME content from 3,381 to 3,876 (mean 3,700 ± 118.7) kcal/kg DM. Thus, the DDGS samples selected for use in this study represented a considerable range in nutrient and energy contents and provided a representative sample of the DDGS materials available to the industry. The prediction equations developed in this study should, therefore, apply to DDGS samples from ethanol plants in the Midwest that are currently supplying product to the swine industry.

In general, correlations between chemical composition components and the DE and ME content of DDGS were relatively weak and, also, differed between the two laboratories. For Laboratory 1, the strongest correlations with DE and ME were for ADF (-0.51 and -0.50, respectively). In contrast, the strongest correlations for DE and ME for Laboratory 2 were with crude fat (0.60 and 0.67, respectively) and crude fiber (-0.56 and -0.52, respectively). Equations to predict the ME of DDGS based on chemical components also differed between laboratories. Equations based on proximate analysis components (ash, crude protein, crude fat, and crude fiber), either individually or in any combination, were relatively poor predictors of ME content. Adding other components to the 4-variable equation based on proximate analysis components, particularly ADF, NDF, and GE, to create 6- or 7-variable equations improved the accuracy of prediction of the ME of DDGS. There was little increase in the accuracy of prediction for equations with more than 7 variables. Equations were developed for the chemical components other than proximate analysis; for these, the 3-variable equation that explained the greatest variation in the ME content of DDGS was based on ADF, NDF, and GE for both laboratories.

This study clearly highlighted the very large variation that is found in practice in the chemical composition and energy content of DGGS samples from different sources in the Midwest of the US. In addition, equations have been developed to predict the ME content of DDGS based on chemical composition. A critically important finding is that these equations differed between the two laboratories used for the chemical analysis. As in many situations, the choice of the most appropriate equation to use will be based on a balance between the accuracy of the equation compared to the costs of carrying out the chemical analyses. Equations based on all possible combinations of the chemical components determined in this study are presented in this report to allow individuals to choose the equation that is most appropriate for the particular situation. These prediction equations relate to ground DDGS samples with a particle size within the range 265 to 403. However, it is important that these equations are fully validated before widespread application can be recommended.

For Further Information contact:
Dr. Mike Ellis
1207 West Gregory Dr.
216 ASL
Urbana-Champaign, 61820
Office: 217.333-6455
[email protected]