The goal of feed formulation is to meet the nutritional needs of the target animal in the most profitable manner possible. This takes into account not only performance, but also optimization of efficiency and in some locations includes environmental nutrient management. Under formulating places the performance at risk, while over formulation is damaging to the costs structure and results in more waste production. Most formulators supply a slight over formulation either by elevating nutrient requirements, or by pegging production slightly higher than actual.
Hindsight really is 20/20. Formulation models are produced from retrospective data; thus the developers have the advantage of knowing the results of feeding programs and working backwards to develop robust equations. To get accurate results, users of the models must make assumptions, such as predicting environmental conditions, the production potential of the animals and use accurate nutrient values for ingredients.
Forages are the most variable part of the feeding program for ruminants. Forage analysis laboratories provide analyses in great detail and are progressive contributors to the feed industry. Laboratory analyses provide on the spot estimations of protein fractions and digestibility of carbohydrate components for forages.
Results from forage analyses often pertain to the enterprise from which they were grown or for which they were purchased, and the forages remain there until consumed. Some concentrate ingredients, such as corn, barley or beans may likewise be produced on a farm and used there. However, it is more likely that these ingredients arrive as commodities, or are components of purchased mixes, and values are estimated from databases. In a feed mill situation, the ingredients may be gone before analytical results are returned from the laboratory.
While less variable than forages, the concentrate and byproduct ingredients warrant consideration and are important to formulation outcome.
Variability in proximate composition
As processing byproducts, many ingredients are by nature somewhat inconsistent. Processing methods can be unique to individual processing plants, and as a result, the composition of the byproduct can be unique to that plant. Likewise, because growing conditions vary from region to region, there can be differences in the base ingredient entering a plant from one region to the next. For example, corn distillers’ grains are noted for their variability. Researchers have found that much of the variability could be associated with the source of the product. Similar plant-to-plant differences in nutrient profile have been noted for soybean products and canola meal.
Within concentrate and byproduct ingredients, some nutrients are less consistent than others. Table 1 shows the variability of several nutrients for some common ingredients. Variation by nutrient needs to be related to relative importance for formulation.
Variability in intrinsic components
While proximate components are readily measured, some variables, important for formulation are obtained from tables. Inconsistency in these values can be caused by many of the same factors as those that affect proximal composition. This would apply particularly to nutrients such as vitamins, micro minerals and amino acids. However, the methods used to determine other components, such as protein and carbohydrate fractions and digestion rates can have analytical biases associated with them.
Table 2 shows analytical biases between two methods used to determine rumen undegraded protein (RUP). Studies found quite large differences between RUP determined by in situ vs. in vitro methods. Furthermore, there was no consistent relationship or pattern in the results.
Another issue arises with the use of database information that may not clearly describe the product at hand. Continuing with RUP as an example, differences in ingredients assumed to be the most consistent can in fact be quite different. Soybean meal is assumed to be quite consistent. In an extensive review, soybean meal, which tends to have very little RUP unless it is heat or chemically treated, can have a considerable range in values within a processing method. Likewise, the RUP value for canola could be related to particular crushing plants.
Variability in digestibility
It is possible for ingredients to have very consistent nutrient specifications, but with digestibility an unseen threat to the prediction of feeding value. The grain considered to be the most consistent in North America is corn (zea mais). Research conducted at the University of Wisconsin (Hoffman and Shaver, 2011) showed that the ability of rumen microbes to digest starch depends on the relative portion of floury as opposed to vitreous starch. Vitreous starch is closely associated with insoluble storage proteins. The study showed that approximately 50 percent of the starch in corn was digested in 14 hours when the corn contained 25 percent vitreous starch. In contrast, only 20 percent was ruminally digested when the corn contained 66 percent vitreous starch. Starch in high-moisture corn becomes more available with ensiling in association with fermentation of the protein associated with it.
Table 3 provides digestibility values for the RUP fraction of selected ingredients. The range in digestibility is quite narrow for solvent extracted soybean meal. In contrast, the range in digestibility for batch-dried blood meal is extensive: from 29 percent to 86 percent of the RUP fraction. The data likewise show that processing methods have an impact on digestibility. Based upon these data, expeller soybean meal is more variable than solvent extracted. Ring-dried blood meal is less variable than batch dried.
Dealing with variability
The models available to nutritionists are in general terms reasonably accurate predictors of nutrient requirements. Accurate information concerning ingredient composition is fundamental to applying the model appropriately. Some steps that can be taken include: • Analyzing incoming ingredients to the extent practical. Ingredients with low variability require fewer analyses than ones that are highly variable. • Knowing, if possible, the source of the ingredient. This is particularly important for byproducts, where processing can have an impact on the value of the ingredient. If possible, source byproducts from the same processors. • Using the results to create your own database. This provides a starting point for ration balancing and ingredient selection. •When balancing for nutrients obtained from file values only, be aware of the method used to obtain the information. For example, if amino acid values can be obtained from one source, any bias should be similar for all ingredients. • Being aware of drifts in the nutrient values of ingredients. Many databases represent years of information. In the interim, processing methods, varieties of ingredients, etc. may have changed. • A change in analytical methods or laboratories may also result in a sudden shift in one or more nutrient values. If no concomitant change is made on the requirement side, then performance variations may result.
Tools are available to allow formulators to take into account ingredient variability in ration formulation. Such tools provide the formulator with the ability to reduce safety margins set for nutrient requirements, and to use and make better use of the ingredients available for ration balancing.