A new method to estimate residual feed intake in dairy cattle using time series data

Area of Science:

• Animal Science
• Dairy Production
• Quantitative Genetics

Background:

• Traditional residual feed intake (RFI) measurement in dairy cows uses linear regression, which overlooks temporal dynamics of feed efficiency components.
• This limitation leads to approximations in RFI results, necessitating a more dynamic approach for accurate assessment.

Purpose of the Study:

• To introduce and validate a novel multitrait random regression model for dynamic feed efficiency assessment in dairy cows.
• To investigate the temporal correlations between milk production, live weight, dry matter intake (DMI), and body condition score (BCS) throughout lactation.

Main Methods:

• A multitrait random regression model was employed to analyze dynamic relationships between key production and intake traits.
• Matrix regression on variance-covariance matrices and animal effects estimated predicted intake, with RFI derived from the difference between actual and predicted intake.
• The model was validated using historical data from 1,469 lactations of 740 cows at Aarhus University.

Main Results:

• High positive correlations were observed between milk/DMI and weight/DMI, peaking mid-lactation.
• Weight and BCS correlations remained stable (~0.4), while milk and weight, DMI and BCS, and milk and BCS correlations decreased over lactation.
• Estimated RFI exhibited classical RFI characteristics (mean zero, phenotypic independence) and showed strong correlations with averaged RFI over lactation.

Conclusions:

• The proposed dynamic model accurately captures temporal variations in feed efficiency components, overcoming limitations of traditional RFI methods.
• This approach provides a more precise estimation of feed efficiency, adaptable for genetic and genomic selection in dairy cattle.
• The dynamic RFI estimation is robust, handles missing data, and offers improved insights into feed utilization across lactation.