Feed accounts for more than 50% of the cost of pig production. From technical and economical points of view, it is therefore essential to know accurately both animals' nutritional requirements and the nutritive values of feeds. A new software tool can help predict the nutritional values say Jean Noblet and colleagues Alain Valancogne, Gilles Tran and Yvan Primot.
By: Jean Noblet, Alain Valancogne, Gilles Tran & Yvan Primot
Pig feed must satisfy requirements for protein (amino acids), minerals, vitamins and energy - of which the latter represents the largest fraction of feed costs. Except for the regions that use very simple diets (i.e. corn and soybean meal alone), compound feed and concentrates can be based on many different raw materials. These feedstuffs have chemical characteristics that vary over time and between regions and they also have highly variable nutritional values. It is thus necessary to characterise them as precisely as possible in order to estimate their values for animals. However, the in vivo determination of nutritional values is complex, time-consuming and costly: feed companies estimate the values rather than measure them. The estimation methods are more or less sophisticated and accurate; often, only the average nutritional values available in feed tables are used, and the effect of chemical composition on nutritional value is not taken into consideration.
In the case of pigs, popular feed tables include for example the one published by INRA and AFZ (Sauvant et al., 2002; 2004; Noblet et al., 2002; 2004) and the Dutch CVB Feed Table. The INRA table lists about 100 raw materials and provides information for the main chemical composition criteria and for energy, protein and mineral values. The energy value is estimated on the basis of the digestible (DE), metabolisable (ME) or net energy contents (NE) with values depending on the pig physiological stage (Figure 1); different values are proposed for the grower pigs (piglets, growing and fattening pigs, young breeders) and for the adult animals (gestating or lactating sows, boars). The studies on which these proposals are based have been summarised in review articles (Noblet et al., 2003; Noblet and van Milgen, 2004; Noblet, 2005).
The comparison of DE, ME or NE systems as estimators of feed energy value and the response of pigs show the clear superiority of the NE system. The protein value is estimated on the basis of the amino acids content and their digestibilities. Several digestibility or availability concepts are potentially usable. However, the standardised ileal digestibility concept is the most commonly used and these digestibility values have been obtained for a large number of raw materials; it allows a satisfying classification of the feed protein value (Sève, 1994; Stein et al., 2007). Contrary to energy, a unique protein value is provided for all stages of pig production. The mineral value of raw materials is assessed through macro and trace elements. In the case of phosphorus through digestibility and availability values. The effect of endogenous or exogenous phytase is also considered for phosphorus (Noblet et al., 2004; Jondreville and Dourmad, 2005). Technological effects (e.g. pelleting and extrusion) are not taken into account in the INRA-AFZ tables and other tables, even though they can affect the nutritional value, especially the energy value of fat-rich feedstuffs such as maize, full-fat rapeseed, or linseed (Noblet, 2005; Noblet et al., 2007).
New software tool
Feed tables provide average nutritional values based on average chemical characteristics that were determined at the time of the publication: such values are not usually applicable to raw materials that have a different composition from the one in the tables. In order to overcome this limitation, INRA, AFZ and Ajinomoto Eurolysine have developed the EvaPig® software.
The main function of the software is the estimation of energy (DE, ME, NE), protein (digestible amino acids) and mineral (digestible phosphorus) values for the raw materials listed in the INRA-AFZ tables (called “reference ingredients” hereafter) according to their actual chemical composition. When calculating the energy value or amino acid and phosphorus digestibility of an ingredient that derives from a reference ingredient, the software uses for the new ingredient a combination of generic equations and specific equations. For instance, gross energy is predicted using a generic equation requiring coefficients for protein, fat and ash that are identical for all ingredients; on the other hand, energy digestibility is predicted from fibre content with coefficients specific to the ingredient (botanical species, type of process, etc). The calculations consist in combining the nutritional value of the “reference” ingredient and coefficients that are applied to the differences in chemical composition between the “new” ingredient and the reference ingredient. The formula takes the following general form:
Y New = Y Reference + a x (X New – X Reference) + b x (Z New – Z Reference) + … where Y is the predicted value and X, Z, etc. are the predictors.
Effect oftechnological processes
The energy values provided in the INRA-AFZ tables refer to mashed ingredients. However, technological processes such as intensive grinding, pelleting, extrusion, or the addition of substances such as enzymes, are known to increase energy digestibility. The increase depends on the ingredient and on the nature of the process. It is mostly noticeable in the growing pig and is supposed to be less important in the adult pig (literature data are not yet available). Therefore, the software tool applies an “energy bonus” that can add up to 5% to the energy values. This bonus can also be used when the energy value of the reference ingredient seems underestimated. Likewise, a negative value can be used when the reference data seems overestimated. This correction is applied to the DE value and consequently modifies the ME and NE values. Similarly, it is possible to use an “energy bonus” for diets (positive bonus only) that have undergone a technological process. For instance, pelleting increases digestibility by 1 to 3 points for diets based on wheat, maize or soybean meal.
The second function of the software is the estimation of energy, protein and mineral values for raw materials absent from INRA-AFZ tables and for complete diets, using only chemical criteria and generic equations (and coefficients). For energy values, the calculation process predicts gross energy, energy digestibility, ME/DE and NE/ME using generic equations. The tool also provides default digestibility coefficients for amino acids that can be changed by the user. However, while they are of great practical interest, generic equations are less precise and fail to take into account ingredient-specific effects such as anti-nutritional factors or the structure of cell walls. For that reason, ingredients and diets should be created using chemical composition only when no other option is available. The third function of the software is the calculation of the nutritional value of diets from the feedstuffs composition. For diets created from a list of calculation ingredients, the chemical and nutritional values are calculated as the weighed contributions of the ingredients, taking into account their incorporation rates and dry matter values. When a nutrient is missing from an ingredient, it will not be part of the diet calculations: for instance, if an ingredient has no net energy value attached, net energy will not be calculated for any diet including this ingredient. In the case of calculation of phosphorus digestibility, the following options are proposed:
• When no phytase is added, the calculation consists of a summation of the contributions of each ingredient, taking into account the incorporation rate and whether or not the diet is processed. For an unprocessed diet, the values will be those of the mash ingredient while for a processed diet, the values will be those of the pelleted ingredients.
• When exogenous phytase is added and the diet is in mash form (i.e. unprocessed), some ingredients contribute to phosphorus release according to their endogenous phytase. Because the effect of phytase on phosphorus release is not linear, the total effect of phytase (endogenous and exogenous) needs to be estimated first, and several calculation steps are necessary to estimate the amount of the released phosphorus due to exogenous phytase, which is then added to the digestible phosphorus of the diet.
• When phytase is added and the diet is in pellet form (i.e. processed), only exogenous phytase contributes to phosphorus release. The released phosphorus is estimated and added to the digestible phosphorus of the diet.
It can be concluded that this new software tool makes it easy to calculate the nutritional value of any raw material suitable for pig feeding. Users can also investigate the changes in raw materials ranking and usage induced by nutritional concepts of varying complexity (for instance, NE vs. DE or digestible lysine vs. total lysine). The software was designed to incorporate, on a regular basis, the latest data on raw materials, so that users can take advantage of the increasing diversification of raw materials (by-products) within an environment where the various feed and food sectors (animal species, humans, industries) compete for dietary resources. <-
EvaPig is free and available in numerous languages. More information can be found on www.evapig.com
Alain Valancogne works at INRA, Gilles Tran works at AFZ and Yvan Primot works at Ajinomoto Eurolysine
Source: Feed Mix Magazine. Volume 17. No. 3