Amino acid balancing in dairy rations: lysine, methionine and everything after
Published 22 March 2026 · Updated 5 July 2026
A dairy cow eating a maize-silage-and-soybean ration does not lack protein in any gross sense. What she lacks, fairly often, are two specific amino acids in the right quantities at the right place: lysine and methionine, absorbed from the small intestine after the rumen has had its say. The modern effort to balance these two, and increasingly a handful of others, has changed dairy formulation from a crude-protein exercise into something closer to the amino acid accounting that poultry nutritionists have done for decades. The change has been slow, partly because the rumen sits in the way and makes everything harder to predict.
Why lysine and methionine come first
A first-limiting amino acid is the one that runs out before the others when the animal is assembling milk protein. Whatever quantity of it reaches the small intestine sets the ceiling; the remaining amino acids are present in surplus and their nitrogen is stripped off and excreted. In maize-based dairy rations methionine is almost always first-limiting, because maize protein is poor in it. In grass-silage diets lysine tends to be the tighter constraint, because the microbial protein made in the rumen on such diets is adequate in methionine but marginal in lysine when yields are high.
The practical target that most formulation systems converge on is a lysine-to-methionine ratio in metabolisable protein somewhere near 3:1, with lysine at around 7 per cent of metabolisable protein and methionine around 2.3 to 2.5 per cent. These numbers have been refined over the past twenty years and they are still argued about at the margins, but the broad shape holds. When the ratio is badly wrong the cow wastes nitrogen, milk protein yield drops, and the nitrogen she excretes ends up in the urine rather than in the bulk tank.
What the rumen does to the plan
The difficulty with amino acid balancing in ruminants is that the feed protein the cow swallows is not the protein she absorbs. Most of it is broken down by rumen microbes to ammonia and resynthesised into microbial protein, which has its own amino acid profile regardless of what went in. Microbial protein is reasonably well balanced—it is a decent source of most essential amino acids—but it is made in a fixed quantity that depends on energy supply and rumen conditions, not on the cow's requirements.
The protein that escapes the rumen undegraded, often called bypass protein or rumen-undegradable protein, retains the amino acid composition of the original feedstuff. It is this fraction that the formulator can manipulate directly, and it is here that the identity of the protein source matters. Soybean meal that escapes the rumen contributes lysine; maize gluten meal contributes methionine but very little lysine. Blood meal is rich in both but variable in digestibility. The job of the formulator is to assemble bypass protein from sources whose amino acid profiles, mixed with microbial protein, meet the target.
That job is harder than it sounds, because the proportion of any feedstuff that escapes the rumen depends on passage rate, particle size, rumen pH and a list of other variables that are hard to hold still on a commercial farm. A ration that balances perfectly in the spreadsheet may not balance at the duodenum, and the cow cannot tell you which side of the error she is on.
Rumen-protected amino acids
The commercial response to this problem has been rumen-protected methionine and, to a lesser extent, rumen-protected lysine: crystalline amino acids coated or encapsulated so that they pass through the rumen intact and release in the abomasum or small intestine. The methionine products work well enough that they are standard practice in many herds. The lysine products have had a more uneven history, because protecting lysine from rumen degradation without also reducing its intestinal digestibility has proved harder, and some of the early products delivered less than they claimed.
When rumen-protected methionine is added to a ration that was marginal, the response in milk protein concentration is typically 0.05 to 0.10 percentage points—a modest shift in a number that usually sits between 3.0 and 3.5 per cent, but one that compounds over a lactation and that the milk cheque notices. The protein evaluation systems used in northern Europe now include explicit methionine and lysine predictions, which means the formulation software can price the protected products against protein sources that supply the same amino acids through bypass.
Beyond the first two
Once lysine and methionine are in balance, histidine becomes the next candidate, particularly in grass-silage-based diets in the Nordic countries and the UK. The evidence for histidine limitation is more recent and thinner, and there is no established commercial market for rumen-protected histidine in the way that there is for methionine. Leucine, isoleucine and phenylalanine have been discussed as potential co-limiting amino acids under specific feeding conditions, but none of them has yet generated the same weight of production data.
There is a philosophical question underneath the practical one: how far down the amino acid list is it worth going? Each step requires more data, more analytical precision and more confidence in the model that predicts supply, and at some point the uncertainty in the prediction exceeds the size of the response being chased. That point is probably closer than the advocates of precision nutrition would prefer.
What amino acid balancing changes in practice
The tangible result of getting this right is not so much more milk as cheaper milk. A ration balanced for amino acids can carry less total crude protein—often 15 per cent instead of 17—without losing milk protein yield, because the amino acids that matter are supplied in the right proportions and less nitrogen is wasted. The crude protein that comes out of the ration is the expensive part: soybean meal, canola meal, protected amino acids. The nitrogen that comes out of the cow is the environmental part: urea in urine, ammonia from slurry. Both go down when the ration is tighter.
In herds that have moved to this kind of formulation the feed cost per litre tends to fall, the nitrogen efficiency of the herd tends to rise, and the two together make a case that is easier to argue to a farmer than a case built on either alone. The practical difficulty remains the prediction: knowing what reaches the small intestine, not just what went into the mixer. That prediction depends on a chain of models, each of them uncertain, and the chain is only as strong as its weakest link.