Amino acids

Balancing rations for individual amino acids improves feed efficiency. Protein production by the cow is limited by the particular amino acid that is in shortest supply in relation to the cow’s requirement for forming amino acid chains. The goal when feeding dairy cows should be to maximize microbial amino acid production as much as possible and then, complement the microbial supply with additional amino acids that are expected to escape rumen fermentation. Amino acid requirements can be expressed using either the factorial method or the ideal protein method.

If we can more accurately predict the nutrient requirements of cows and the nutrients supplied by rations, we should be able to reduce feed costs and reduce nutrient wastage while promoting maximum milk production. Protein is composed of amino acids. We know that cows require certain amounts of individual amino acids for production. Balancing rations for the individual amino acids rather than for crude protein will help us fine-tune rations and improve feed efficiency.

A protein is a chain of 50 or more individual amino acids. There are 20 different amino acids. Chemically, each amino acid contains carbon, nitrogen, oxygen, and hydrogen in different amounts and in different placements. A few amino acids also contain sulfur. In the laboratory analysis for crude protein, the nitrogen content of feed is measured. It is assumed that protein contains 16% nitrogen. The percentage crude protein is calculated by multiplying the percentage of nitrogen in the feed by the factor 6.25 (100/16).

The cow absorbs and uses individual amino acids rather than protein. She does not have a "protein" requirement but rather a different requirement for each of the 10 essential amino acids, including: phenylalanine, valine, threonine, tryptophane, isoleucine, methionine, histidine, arginine, leucine, and lysine. The term "essential" amino acid refers to those amino acids that the cow cannot produce by rearranging the chemical configuration of a different amino acid. Each essential amino acid must be supplied at the intestine from the diet or from the rumen microbes. The cow also requires “nonessential” amino acids but can make them in sufficient amounts to meet needs.

The cow takes individual amino acids and combines them in chains with specific sequences to make protein in the form of milk, a calf, or muscle. The cow's protein production is limited by the particular amino acid that is in shortest supply in relation to her requirement for forming the amino acid chains. That amino acid is called the "first-limiting" amino acid in the diet. The “first-limiting” amino acid of any diet will depend upon the amino acid profile of the ingredients in the ration. For dairy cows, it is usually lysine or methionine. Arginine is also a concern. Three different branched-chain amino acids can also become limiting (valine, isoleucine, and leucine).

Making protein in the cow is like putting together a big puzzle with many different pieces each going together in a certain order. If the cow has two of the same puzzle piece but is missing a different piece, no matter how hard she tries she will not be able to complete the puzzle. If extra methionine is supplied but lysine is actually the "first-limiting" amino acid, a production response will not be seen.

Amino Acid Supply:

The rumen microbes, undegraded dietary amino acids (from UIP or rumen-protected amino acids) and endogenous amino acids (from sloughed off cells and secretions in the digestive tract) all contribute to the amino acid pool flowing out of the rumen and available for absorption at the small intestine. Rumen microbial amino acids actually contribute about 50-75% of the total amino acids. Endogenous contributions are minimal.

The goal when feeding dairy cows should always be to maximize microbial amino acid production as much as possible and then, supplement with additional amino acids which are expected to escape rumen fermentation. This strategy provides as economical source of amino acids and, due to growth of the microbial population, digestibility of starch and fiber is increased to provide more volatile fatty acids for energy.

1. Microbial Amino Acids:

The amino acid profile of microbial protein is very similar to the amino acid profile of milk. This makes microbial protein a "high-quality" protein similar to the animal proteins that we use in feeds. Predicting the amount of microbial amino acids produced in the rumen is difficult. The amount of microbial amino acids produced in the rumen is dependent on carbohydrate availability in the rumen, synchronization of carbohydrate and protein availability, rumen available fat, rumen pH, the rate at which rumen contents turn over, and the types of rumen microbes.

The NRC (2001) predicts microbial protein production from the amount of predicted digested organic matter in the rumen. They use an equation which relates microbial protein yield determined in past studies using intestinally cannulated cows to the adjusted TDN (discounted for fat) content of the diet fed to the cows in each of those studies (Microbial CP (grams) = 130 * Adjusted TDN intake (kg)). If the diet is limiting in ruminally degraded feed protein (DIP or RDP), predicted microbial protein production is limited. When intake of RDP is less than 1.18 times the adjusted-TDN predicted Microbial CP yield, then Microbial CP is calculated as 0.85 times RDP intake. The amount of microbial amino acids can then be predicted based on the assumed amino acid profile of the microbial protein.

There are still problems with the NRC (2001) equations for predicting microbial amino acid flow. T he NRC equation ignores the fact that the amount of RDP that is needed by the microbes is dependent upon the amount of energy available for the microbes to use. Adjusted-TDN is an especially poor predictor of microbial amino acid production at higher energy intakes. There are larger differences in rumen carbohydrate and protein availability with those diets. Also, diets designed for high production usually have more rumen available fat. Because of the higher levels of fermentable carbohydrate in high production diets, they result in lower rumen pH. The extra fat and low pH both negatively impact microbial growth. Rumen turnover rate increases as dry matter intake goes up. As rumen turnover rate increases, the efficiency of rumen microbial growth also increases. The NRC (2001) equations do not adequately account for rumen turnover rate. Products such as yeast culture and other fermentation products can also enhance microbial growth and need to be factored into prediction equations.

In order to improve the prediction of microbial amino acid supply, computerized models have been developed to mathematically describe and compile scientific data. The Net Carbohydrate / Amino Acid Model (and CPM Model) has been developed at Cornell University , the University of Pennsylvania , and Miner Institute. It is being used on farms by some nutritionists. This model classifies the microbes as either fiber or non-fiber digesters. These microbes are predicted to grow and produce microbial protein differently based on predicted available carbohydrate supply, amino acid supply, and rumen pH. Carbohydrate and amino acid supply in the rumen are predicted based on rates of digestion and rates of passage of feeds. Because the Model takes into account more of the factors that affect microbial growth, it has the potential to do a better job than the NRC equations in predicting microbial amino acid supply.

Amino Acid Composition (%) of Body Tissue, Milk, and Microbial True Protein





Lysine   8.2 8.3 10.46


2.7 2.7 2.68


6.8 3.7 6.96





Isoleucine   5.5 6.0 5.88
Leucine   7.2 10.0 7.51

 Source: Mantysaari, P.E., C.J. Sniffen, and J.D. O'Connor. 1989

Researchers estimate that about 70% of microbial protein is composed of amino acids (true protein (Korhoren et al., 2002). The amino acid composition of the microbial true protein is fairly constant and doesn't change much with diet.

2. Rumen Undegraded Dietary Amino Acids:

Rumen undegraded protein (UIP) contributes a significant but variable supply of amino acids to the small intestine. The amino acid profile of the UIP can be very good, such as in animal proteins like fishmeal or blood meal. However, some rations which appear to be adequate in amount of UIP arriving at the small intestine may be marginal in certain amino acids. For example, corn is known to be low in lysine. A high-producing cow's production may be limited by lysine if most of her UIP is supplied from corn distillers' grains or corn gluten meal. A blend of UIP sources which includes some corn, soy, and animal protein should more adequately provide the cow with amino acids in the proportions needed to manufacture milk and meat proteins.

Amino Acid

Good Source

Methionine   Fish meal, Corn Gluten Meal, Rumen Bypass Methionine


Blood Meal, Fishmeal, Processed Soybean, Rumen Bypass Lysine


Feather Meal, Fishmeal, Processed Soybean

The amount of dietary amino acids that the microbes digest is a function of the competition between the rate at which the microbes can digest the protein and the rate at which the protein passes out of the rumen. Many feed companies and common computer feed programs use set degradability values for feeds. These values are obtained by incubating the feed in the rumen (in situ) or in a buffer solution and enzyme for the amount of time that the feed would be expected to be in a cow's rumen (16 h for concentrates and 30 h for forages). But, the true UIP of a feed will not always be the same because the rate of passage of that feed will vary according to the amount of feed the cow eats. The higher the dry matter intake, the faster a feed will pass through the rumen. UIP is most correctly calculated as a function of the rates of degradation and rates of passage.

To predict degraded feed protein (DIP or RDP) and undegraded feed protein (UIP or RUP), the NRC (2001) publication recommends that feed proteins be broken down into 3 fractions (A, B, and C). The A fraction is assumed to be entirely degraded in the rumen. The C fraction is not degraded in the rumen. The amount of B fraction digested in the rumen is a function of its predicted rate of digestion (derived from in situ data) and its predicted rate of passage (estimated based on dry matter intake, percentage dietary concentrate, percentage dietary NDF and feed moisture content). The digestibility of undegraded feed protein varies by feed source in the NRC (2001) publication based on past research data.

In the Net Carbohydrate / Amino Acid Model, proteins are fractionated into 4 rumen available fractions. Rates of degradation are assigned to each of the four fractions and rates of passage are assigned to each feed. UIP is calculated as a function of the rates of degradation and rates of passage of each protein fraction.

The rumen microbial population varies with the diet being fed. This can affect rumen protein degradation and estimates of bypass amino acids. In a study conducted at Miner Institute (de Ondarza and Sniffen, 2000), an early lactation cow (60 DIM) and a late lactation cow (142 DIM) were fed diets expected to produce different microbial masses and rumen environments. The amount of protein removed from feeds by the rumen microbes in eight hours was determined using the in situ (dacron bag) procedure. Differences in rumen degradability of proteins were found between diets. For example, when roasted soybeans were degraded in the late lactation cow, about 65% of the CP disappeared. But in the early lactation cow consuming a higher ratio of concentrate to forage, an average of 75% of the CP disappeared. These differences were statistically significant but the differences were not consistent among different feeds analyzed.

The amino acid profile of the undigested dietary protein will not be the same as that of total protein consumed. The microbes preferentially use some amino acids and leave a higher percentage of other amino acids undigested. Unfortunately, this preferential usage of amino acids is difficult to predict. The table below shows the ratio of amino acid concentrations in soybean meal and in fishmeal after and before rumen incubation. A value below 1.00 indicates the amino acid was preferentially degraded in the rumen. A value above 1.00 indicates the amino acid was not degraded in the rumen as much as other amino acids in the feed.

Ratio of Amino Acid Concentrations After and Before Rumen Incubation


Soybean Meal  



0.90 1.11


0.92 0.70


0.88 0.94


1.01 1.11


1.02 1.15

Source: Susmel et al., 1989

Dr. Peter Robinson, a cooperative extension specialist with the University of California , has 3 papers on his website which contain amino acid data on rumen escape protein.

Rumen-Protected Amino Acids
Rumen-protected amino acids can elicit a milk production response only if they are digested post-ruminally and only if they supply “limiting amino acids” to the cow. Of course, in order to predict needs for specific amino acids, an accurate rumen model needs to be used for ration balancing.

Absorption and Efficiency of Amino Acid Usage:

The amino acids must be absorbed from the intestine into the bloodstream. Amino acids are not all absorbed with the same efficiency. They sometimes compete for absorption sites. Researchers have also found that peptides (small chains of amino acids) are sometime s absorbed faster than individual amino acids. The amino acids in the bloodstream are used with different efficiencies depending on the function that they are used for. As an example, 1 gram of lysine may arrive at the small intestine of a cow. If 80% of that lysine was digested and absorbed out of the intestine, there would be 0.80 g of lysine available in the bloodstream. If the lysine in the blood were used for production of a calf with an efficiency of 85%, one would have 0.68 g of lysine (0.80 * 0.85) available to meet the cow's lysine requirement.

Use of Individual Absorbed Amino Acids for Different Functions

  Maintenance Growth Reproduction Milk


0.85 0.85 0.85 0.88


0.85 0.85 0.85 0.98


0.85 0.85 0.66 0.42


0.66 0.85 0.66 0.72


0.66 0.85 0.66 0.62






 Source: Evans and Patterson, 1985

Amino Acid Requirements:

Dairy cows require amino acids to make milk and muscle protein, to make the protein in a growing fetus, and to make the proteins they need to maintain themselves (such as enzymes required to digest feeds). Each of these proteins is made up of a different profile of amino acids. For this reason, a cow will require different amounts of each of the essential amino acids depending on her stage of lactation, growth, and pregnancy.

Amino acid requirements can be expressed using either the factorial method or the ideal protein method. With the factorial method, requirements are based on the needs of the animal (for example, the amount of milk to be produced and its amino acid content) and the various efficiencies with which amino acids are absorbed and utilized. Unfortunately, these efficiencies are difficult to estimate. The mammary gland takes up more of some amino acids than it puts out in milk. With some non-essential amino acids, the mammary gland takes up fewer than are found in the milk produced. The pattern of amino acids needed for milk production is therefore different than the amino acid pattern of milk.

Uptake of amino acids by the mammary gland versus output of amino acids in milk (g/100 g amino acids)

Amino Acid 

Total Uptake Mammary Output








8.80 5.79




Lysine  9.14 3.40


2.82 2.71


4.51 4.75


4.76 3.72


10.01 5.89

Evans, 1999

Body cells use active transport mechanisms to take up amino acids. There are 3 different types of transport for amino acids: neutral amino acids (threonine, leucine, valine, isoleucine, phenylalanine, methionine, cystine/cysteine, and tryptophan), basic amino acids (histidine, arginine, and lysine), and acidic amino acids. Within an amino acid type, one amino acid can compete for and inhibit the transport of another amino acid. Because of the potential negative effects of amino acids on each other, the factorial method may overestimate production responses when there is an excess of one amino acid. For example, in one study where lysine was the first-limiting amino acid and extra methionine was supplemented, milk protein production was actually reduced (Rulquin and Verite, 1993).

For this reason, the ideal protein method (ratio approach) is more commonly relied upon to balance rations for amino acids. The ideal protein method is based on research study responses to amino acids expressed as a percentage of either metabolizable protein or intestinal essential amino acids. The relationship of amino acids to each other drives the ration rather than the actual amount of individual amino acids absorbed. According to the research of Schwab at the University of New Hampshire , 5% of the intestinal essential amino acids should be methionine and 15% should be lysine. According to research from Rulquin and Verite in France , 2.5% of the metabolizable amino acids should be methionine and 7.3% should be lysine. The NRC (2001) publication summarized a large dataset and concluded that 2.2% of the metabolizable amino acids should be methionine and 7.2% should be lysine. Without supplementation of individual amino acids, it's hard to reach these goals. According to Sniffen et al., 2001, milk protein is reduced when methionine is less than 2.1-2.2% of metabolizable amino acids and lysine is less than 6.0-6.5% of metabolizable amino acids. Sniffen et al., 2001 analyzed the results from 21 experiments that were conducted with rumen-protected lysine and methionine and a range of concentrates and forages. For milk production and milk crude protein yield, the optimum amount of methionine was 2.07% of metabolizable protein and the optimum amount of lysine was 7.04% of metabolizable protein.

Based on their modeling work, Sniffen et al., 2001 endorsed the ratio approach to amino acid balancing. However, it was recognized that the metabolizable protein requirement must still be met and that factorial supplies of individual amino acids should still exceed factorial requirements. At this time, computerized models using both the factorial approach and the ideal ratio approach for balancing amino acids should be the most effective.


Chalupa, W., R. Boston, C.J. Sniffen, and D.G. Fox. 1998. Development of dairy cow nutritional models. In: Proceedings of the Advanced Dairy Workshop, Syracuse, NY.

de Ondarza, M.B. and C.J. Sniffen. 2000. Diet affects rates of ruminal protein degradation. Feedstuffs. May 8, 2000, p. 14.

Evans, E.H. and R.J. Patterson. 1985. Use of dynamic modelling seen as good way to formulate crude protein, amino acid requirements for cattle diets. Feedstuffs. October 14, 1985, p. 24.

Mantysaari, P.E., C.J. Sniffen, and J.D. O’Connor. 1989. Application model provides means to balance amino acids for dairy cattle. Feedstuffs. May 15, 1989, p. 13.

National Research Council. 1989. Nutrient requirements for dairy cattle. 6th rev. ed. update 1989. Natl. Acad. Sci., Washington, DC

Sniffen, C.J., R.W. Beverly, C.S. Mooney, M.B. Roe, and A.L. Skidmore. 1993. Nutrient requirements versus supply in the dairy cow: Strategies to account for variability. J. Dairy Sci. 76:3160.

Susmel, P., B. Stefanon, C.R. Mills, and M. Candido. 1989. Change in amino acid composition of different protein sources after rumen incubation. Anim. Prod. 49:375.

Evans, E. 1999. Criteria for assessing amino acid utilization for milk protein synthesis. Proc. Eastern Nutr . Conf., p. 113-123.

Evans, E. 2003. Practicalities of balancing diets for amino acids. In: Proceedings of the Tri-State Dairy Nutrition Conference, Fort Wayne , Indiana , p. 133.

Korhoren, M., S. Ahvenjarvi, A. Vanhatalo, and P. Huhtanen. 2002. Supplementing barley or rapeseed meal to dairy cows fed grass-red clover silage. II. Amino acid profile of microbial fractions. J. Anim. Sci. 80:2188.

National Research Council. 2001. Nutrient requirements for dairy cattle. 7 th rev. ed. Natl. Acad. Sci., Washington , DC

Rulquin, H. and R. Verite. 1993. Amino acid nutrition of dairy cows: production effects and animal requirements. Page 55 in Recent Advances in Animal Nutrition. P.C. Garnsworthy and D.J.A. Cole, eds. Nottingham University Press.

Rulquin, H., P.M. Pisulewski, R. Verite, and J. Guinard. 1993. Milk production and composition as a function of postruminal lysine and methionine supply: a nutrient-response approach. Livest. Prod. Sci. 37:69.

Rulquin, H., R. Verite, J. Guinard, and P.M. Pisulewski. 1995. Dairy cows requirements for amino acids. Page 143 in Animal Science Research and Development: Moving Toward a New Century. M. Ivan, ed., ISBN 0-662-23589-4, Centre for Food and Animal Research, Ottawa , Canada .

Sniffen, C.J., W.H. Chalupa, T. Ueda, H. Suzuki, I. Shinzato, T. Fujieda, W. Julien, L. Rode, P. Robinson, J. Harrison, A. Freeden, and J. Nocek. 2001. Amino acid nutrition of the lactation cow. In: Proceedings of the 2001 Cornell Nutrition Conference for Feed Manufacturers, Rochester , NY , p. 188.

Related Links:

Using Protected Amino Acids


Mary Beth de Ondarza

Mary Beth de Ondarza
45 articles

Nutritional consultant for the dairy feed industry at Paradox Nutrition, LLC.

Look to Paradox Nutrition, LLC for providing:

  • Controlled research studies on commercial dairy farms
  • Computer modeling with complete nutritional assessment and statistical analysis
  • Complete literature review of a nutrition topic
  • Nutrition writing for company technical bulletins or popular press
  • Dairy nutrition seminars with new, practical, and understandable information for consultants and dairy producers
  • On-farm nutritional consultation

Dr. de Ondarza received her Ph. D. from Michigan State University and her Masters Degree from Cornell University, both in the field of Dairy Nutrition.

Read more »

Paradox Nutrition

Paradox Nutrition

Paradox Nutrition, LLC is a nutritional consultation business for the dairy feed industry. Mary Beth de Ondarza, Ph.D. is the sole proprietor.

Read more »