Page Banner

United States Department of Agriculture

Agricultural Research Service


Location: Location not imported yet.

Title: Data Collection Ratings and Best Prediction of Lactation Yields)

item Cole, John

Submitted to: International Committee on Animal Recording(ICAR)
Publication Type: Proceedings
Publication Acceptance Date: 6/16/2008
Publication Date: 1/12/2009
Citation: Cole, J.B. 2009. Data Collection Ratings and Best Prediction of Lactation Yields. International Committee on Animal Recording (ICAR). ICAR Tech Ser. 13:403–406. 2009.

Interpretive Summary: Best prediction is a flexible tool for accurately modelling milk, fat, and protein yields and SCS in lactations of any length. Data collection ratings allow US dairy producers more choices of sampling programs than those in many other countries, and almost all data from the field can be used for genetic evaluation. The program provides lactation-to-date, 305-d, 365-d, and projected yields, as well as BP of daily yields. The software is in the public domain and may be downloaded from the AIPL website at:

Technical Abstract: Best prediction (BP) has been used in the US to calculate lactation yields of milk, fat, protein, and somatic cell score (SCS) from test day data since 1998, when it replaced the test interval method (TIM) used since 1969. It is more complex than TIM but also more accurate, particularly when testing is less frequent.The BP software was recently modified to use breed-, parity, and trait-specific lactation curves, accommodate lactations longer than 305-d, and provide additional output, such as daily yields. Data collection ratings (DCR), which describe the accuracy of lactation records from a variety of test plans, are calculated as a function of each trait's reliability. Predictions of daily yield were validated using daily milk weights from university research herds. Data collection ratings for fat and protein increased when multiple trait BP was used. Many cows can produce pro'tably for >305 DIM, and the revised BP program provides a flexible tool to model records of any reasonable length and testing pattern.

Last Modified: 05/24/2017
Footer Content Back to Top of Page