LET’S start with a question. Is a modern nutrition model like a spreadsheet with rows of ingredients and columns of nutrients describing the details of that ingredient from which a dairy cow can make milk?
I could build a spreadsheet like this in an afternoon. To connect it to the animal I plan to feed, I could do a quick web search to find out how many grams, pounds, or calories of various nutrients I would need to support various levels of growth and milk production.
With some basic Excel formulas, I could enter diets with actual pounds and then, using weighted averages of the combined nutrients of the diet entered, compare it to what an animal needs to make a certain level of milk. I could even make the “results” column indicate a red or green number showing how the ration will supply each nutrient compared to the requirements.
Even if we made it a bit more complicated and added ingredient cost, how similar is this to the modern nutrition models using Cornell Net Carbohydrate and Protein System (CNCPS), National Academies of Sciences Engineering and Medicine (NASEM), or something similar to build a modern dairy ration?
Though modern nutrition models do have rows of ingredients and columns of nutrients, this is just the starting point for a much more complex process happening behind the scenes.
Why does this matter to a nutritionist or a dairy producer? Even if a CNCPS- or NASEM- based nutrition model is not simply rows, columns, and weighted averages, why does this matter?
This topic, but in much more detail, was the subject of a full morning session at the recent annual meeting of the American Dairy Science Association (ADSA) in Milwaukee, Wis. As I summarize the different sessions, it will become clear that a spreadsheet is not a good way to think of the current models.
All in the details
To start off the morning, Andrew LaPierre offered a detailed history of the CNCPS. Since it is either the heart of, or at least a significant part of, most modern models, it is helpful to understand how it came to be. Starting back in the 1970s, nutrition faculty, extension, and research leaders understood the exact shortcomings of using weighted averages to feed animals, and began to build a better system.
After many iterations, a few name changes, and a new generation of leaders, CNCPS is now at the forefront of dairy ration formulation, and is used in one way or another to build the majority of dairy diets in the U.S. and around the globe.
A complex model is necessary to describe the net effect: The model takes the complicated biological process that starts with chewing and ends with mostly milk, feces, and urine, and it attempts to quantify only the parts of this digestive and metabolic process that are useful for creating growth, fetal development, and lactation. Think of taking your gross income and subtract things like taxes and retirement savings to then arrive at a net dollar amount that ends up being direct-deposited into your bank account for use in your cost of living.
The second key point is that weighted averages fall short since every unique blend of ingredients in every unique diet interact in different ways. Various principles of these interactions are built into the model’s equations to predict how this blend will create the net nutrients the cow can use for production. These equations are part of the “biology” of the model.
The most important thing to remember is that everything impacts everything else. It is a tangled web of interlocking principles, as opposed to simple weighted averages of nutrients.
In the paycheck analogy, it would be as if checks had ever-changing local, state, and federal taxes, with each change impacting every other tax and retirement savings deductions. All of this together impacted the net amount that hits your bank account.
Dairy’s diminishing returns
The third key point is that the efficiency with which a cow converts nutrients to milk is nonlinear. The best way to think of this is the commonly understood law of diminishing returns. Let’s use a student cramming for an exam as an example. The first hour or twoof studying the information will likely yield good results when taking the test. Each hour of additional study the day before a test will likely be met by fatigue and boredom, and a reduced retention of more information.
In modern nutrition models, simple weighted average ration formulation assumes that each incremental nutrient added to the diet will create milk solids at the same rate. In reality, most nutrients follow a reducing response curve. As you add more, the response with each increment is likely reduced — things begin to go backwards.
Focus on fat
The second session was led by Jonas de Souza. As a thought-leader in the topic of feeding fatty acids, de Souza described how various levels and different blends of commercially available fat sources have complicated result predictions. He described how various fatty acids interact with each other and other nutrients in the diet, making responses difficult to predict.
Current commercial modeling does not yet include these complex interactions, requiring the formulator to consider these expectations outside the actual model biology. Improved fat models will be available at some point in the future, taking out some of the complexity and economic risk of adding more fat to diets.
Protein and beyond
The next session on amino acids and protein was presented by Veridiana Daley. Having been included in some of the amino acids and protein portions of the NASEM, Daley was an excellent choice for this session. Interactions in the rumen and variable efficiencies of the transfer of dietary protein to milk protein contribute to the complicated task of balancing protein and amino acids in diets.
Ration models using CNCPS and NASEM principles allow the formulating nutritionist to best build the protein side of the diet. Just like adding fatty acids, balancing for amino acids will likely add cost, and it will be necessary to accurately predict a milk income response. Efficiency of transfer of dietary protein to milk protein will continue to drive research and modeling. Until the picture is clearer, the experienced nutritionist needs to communicate these expense risk realities to the producer and make a plan with some level of uncertainty.
An interesting slide in this presentation showed the biological variation in response to increased metabolizable protein (MP). In a study completed at Virginia Tech University, 47% of cows had a linear response to increased MP, 31% had a quadratic response (think diminishing returns), and 13% showed no response at all. This certainly has an impact on how we feed groups of cows on the same dairy.
The session was concluded with a panel discussion of consulting nutritionists, of which I was honored to be included, along with Nate Elzinga and Logan Morris. Since the room was filled with industry professionals, academic individuals, and students, questions varied from what exact animal in a pen we formulate for, how we each use the plethora of data available from a modern dairy, and discussions of optimization within the CNCPS-based models. The morning session made it clear that ADSA is adjusting annual meetings to add value to field nutritionists, feed company professionals, and others working to feed cows better every day.
Formulating dairy rations is not for the faint of heart. Until you are feeding actual cows and lie awake at night wondering if the ration you just emailed to be fed tomorrow will work, much of this is a bit academic. I, for one, am appreciative of the many research projects ongoing to help us have more accurate models that allow us to build better rations. I depend daily on these models to best meet the needs of the cows and my clients who own them.
.jpg)
