PROT1 (6-00)
Genetic evaluations for true protein

P.M. VanRaden and R.L. Powell
Animal Improvement Programs Laboratory, ARS-USDA, Beltsville, MD 20705-2350
301-504-8334 (voice) ~ 301-504-8092 (fax) ~ ~

Beginning in August 2000, genetic evaluations for protein in the United States will estimate genetic merit for true protein instead of crude protein.

U.S. protein records

Most U.S. farmers are now paid for true protein rather than crude protein. To be consistent with milk purchasers, the U.S. milk-recording system switched from measuring crude protein in milk to true protein on May 1, 2000 (

True protein is a measure of only the proteins in milk, whereas crude protein is a measure of all sources of nitrogen and includes nonprotein nitrogen, such as urea, which has no food value to humans. This nonprotein nitrogen is a waste product when making cheese and other manufactured products.

Current milk-testing equipment measures peptide bonds, a direct measure of true protein. However, in the past, the machines had been calibrated to report crude protein instead.

The percentage of milk that is true protein is lower than the percentage that is crude protein by an approximate difference of .19%. Because the nonprotein nitrogen of individual milk samples is not measured but assumed to be constant, crude protein records can be converted to true protein with the following formulas:

true protein (%) = crude protein (%) – .19%;
true protein (lb) = crude protein (lb) – [.0019 × milk (lb)].

Breed differences

For breeds with higher crude protein percentages, the .19% of nonprotein nitrogen is a smaller fraction of the total nitrogen. The following table shows average yields, percentages, and ratios of true to crude protein from the first-lactation, mature-equivalent records of cows that were born in 1995 and that had records that were usable for genetic evaluation:

Breed Milk,

Difference, lb
(true – crude)

Ayrshire 16,750 646 3.11 520 552 –32 .942
Brown Swiss 19,269 773 3.29 634 671 –37 .945
Guernsey 15,391 682 3.28 505 534 –29 .945
Holstein 23,346 845 2.95 689 733 –44 .940
Jersey 16,017 737 3.54 567 597 –30 .949
Milking Shorthorn 15,934 565 3.05 487 517 –30 .941

Predicted transmitting ability

Predicted transmitting abilities (PTA's) for true and crude protein percentages are the same because subtraction of a constant from all protein percentages does not affect within-herd deviations for protein percentage. However, the PTA for true protein yield (lb) generally is smaller (closer to 0) than PTA for crude protein yield because a smaller fraction of the milk is considered to be protein. For those few bulls with a positive PTA for milk but a negative PTA for crude protein yield (or vice versa), PTA for true protein yield is slightly greater (farther from 0) than PTA for crude protein yield. Crude protein PTA's can be converted to true protein with the following formulas:

PTA true protein (%) = PTA crude protein (%);
PTA true protein (lb) = PTA crude protein (lb) – [.0019 × PTA milk (lb)].

The standard deviation of PTA for true protein yield (lb) for Holsteins is .940 times the standard deviation of PTA for crude protein yield. The correlation of PTA for true protein yield with PTA for crude protein yield was .9993 for Holstein bulls that were designated as being in active artificial-insemination service for February 2000 USDA-DHIA genetic evaluations. The average PTA protein for those bulls changed from 54 lb for crude protein to 51 lb for true protein, but little reranking occurred. The main result of changing from crude to true protein is to reduce the standard deviation of protein yield (lb) by 5 to 6%.

Other U.S. protein reporting

Supplemental information from USDA, such as the average yields for cows and daughter yield deviations for bulls, also will report true protein. Historical records from breed associations and dairy records processing centers may continue to be reported in the original units of crude protein.

International protein reporting

One concern of the U.S. dairy industry was whether other countries report and pay for true protein or crude protein. A questionnaire was sent from USDA to countries that participate in Interbull Centre (Uppsala, Sweden) services. The survey included questions on what protein measures were used for testing, payment, and genetic evaluation within a country. However, all countries reported that only a single measure was used for all categories as shown in the following table.

Protein measure Countries

Crude protein Austria
Czech Republic
New Zealand
The Netherlands
United Kingdom

True protein Australia France Hungary

Italy responded that some farmers are paid for both the quality and quantity of protein and that bonuses are paid for the B allele for kappa casein. Belgium uses a multiplier of 1.05 to convert true protein yield (kg) from other countries to crude protein yield.

Interbull genetic rankings are not affected by a nation's choice of protein measure. Changes in averages and variation that are related to the use of crude or true protein are accounted for by the equations that are used to combine national data.