MP is not crude protein

Crude protein is a convenient lab number: nitrogen × 6.25. It does not tell you how much true protein the cow absorbs after rumen microbes and bypass protein finish their work. That absorbed supply is closer to what nutritionists call metabolizable protein (MP), and the amino acid profile of that MP is what builds milk protein.

Much of dietary protein is broken down in the rumen. Microbes rebuild amino acids into microbial protein, which later digests in the small intestine. Some feed protein escapes (“bypass”) and also contributes amino acids downstream. The cow’s usable amino acid pool is the sum of those streams, not the crude protein on the tag.

On a feed report, those two streams are rumen-degradable protein (RDP) and rumen-undegradable protein (RUP). Microbes need fermentable starch and fiber to turn RDP into microbial protein (see microbes and fermentation). NASEM (2021) suggests a general minimum of about 10% of dry matter as RDP, because intake can drop when RDP runs short. How much of the RUP is then digested in the small intestine differs by feed.

The RUP line on a feed report is calculated from a degradation rate and an assumed passage rate, and faster passage lets more protein escape (what leaves the rumen). Checked against measured flows, the NRC (2001) model overpredicted RUP in high-RUP feeds, so the 2021 model lowered its passage rates. The same feed now prints less RUP and more RDP, most of all in high-RUP concentrates. Compare RUP values only within one system.

When diets are short on key essential amino acids relative to MP (lysine and methionine are the usual first limits in many North American dairy diets, and histidine can be co-limiting), milk protein yield and nitrogen efficiency suffer even if crude protein looks “high enough.” NASEM (2021) moves the conversation toward essential amino acid supply and response, not a single CP finish line.

Efficiency changed too. The older model assumed the MP used for milk became milk protein at a fixed 67%. The 2021 model dropped that constant: milk protein is predicted mainly from absorbed histidine, isoleucine, leucine, lysine and methionine plus digestible energy, and each extra gram returns less as total supply rises. Efficiency is calculated for the diet and compared with the model's target efficiencies, about 0.70 to 0.75 for these five amino acids. An amino acid running above its target is the likeliest to respond to more supply. One well below target is probably in surplus; the extra is broken down and its nitrogen excreted.

Exact MP and amino acid targets belong with the herd nutritionist and the model they use. The atlas lesson is directional: stop treating crude protein as the finish line. Ask whether the rumen is fed for microbial protein, whether bypass protein is intentional, and whether Lys/Met (and sometimes His) balance is part of the conversation.