LAUSR

The increased availability of precision technologies in dairy production systems coupled with improved computational power has increased interest in individualized feeding. The objective of this study was to determine if small groups of dairy cows with a similar response to dietary adjustment could be identified based on individual animal efficiency changes in response to short-term diet shifts mimicking automated precision feeding. A replicated latin square with 4 groups of 6 cows was used to evaluate responses to feeding increasing amounts of ground corn, soybean meal, chopped grass hay, or total mixed ration for 9 d periods. Throughout each period, treatments were top dressed at incrementally increasing rates to provide 0 to 20 percent of calculated net energy intake. Cows were group housed and fed once daily in Calan gates for 36 days. The feed to milk ratio (FtM) was calculated as feed intake per unit milk yield, so that a lower FtM was more considered efficient. To test the idea that cows could be grouped by their FtM response to different feeds, data were split by top dress type and each day’s FtM was used as a variable for each cow. Principal Component Analysis (PCA) was performed to determine cows that had the most similar FtM behaviors over the 9-day feeding period. After PCA, k-means clustering was performed on the first two dimensions from PCA to discern groups of cows in the data. This approach can inform precision feeding strategies by calculating within-group average MYE and assigning cows to whichever feeding strategy generated the lowest MYE. The fact that cattle segregate into different efficiency groups depending on the supplement type provided suggests that individualized feeding schemes must be designed with individual animal interactions with supplement type in mind.