Visualizing differences in efficiency of beef cattle
Using computer vision and AI
By Derek Brake
There is at least a 15% difference between how well the upper one-third and lower one-third of finishing cattle in the United States use feed to support growth. According to the July Focus on the Feedlot report from Kansas State University that difference in feed efficiency between the most efficient and least efficient cattle entering Kansas feedlots would be nearly $150 per head.
While differences in feed efficiency are noticed every time we look over a ledger when closing out groups of cattle in the feedlot, differences in feed efficiency in beef cattle are not solely limited to the feedlot. In fact, differences in feed efficiency between the most efficient and least efficient pregnant beef cows appear to be even greater than differences in cattle in the feedlot with the upper one-third of cows using only about three-quarters of the amount of feed needed by the lower one-third of cows to sustain themselves and their pregnancy.
Understanding the importance of individual differences in how beef cattle efficiently use feed for production is not new, and the beef industry has made notable improvements in how efficiently beef is produced. Indeed, across the past 45 years, beef production in the United States has increased by over 1.25-billion pounds despite a 22% decrease in the number of cattle harvested for beef.
Yet, individual measures of feed efficiency in cattle have not closely mirrored improvements in overall efficiency of beef production, and a substantial amount of the advances in efficiency of beef production seem to be closely related to increases in carcass weights.
While increases in final body weight at harvest can improve the amount of beef produced relative to the amount of cattle in the U.S. beef herd this strategy does not largely improve the economic or environmental sustainability of beef production. Improvements in the efficient use of feed for growth by beef cattle, however, create opportunities to simultaneously reduce carbon and nitrogen excretions by cattle and reduce feed costs.
Substantial variation in feed efficiency between growing beef cattle remains among the largest barriers to improving individual feed efficiency, and an inability to accurately identify individual differences in potential feed efficiency between beef cattle prevents producers from placing cattle into production systems that could optimize economic returns and minimize nutrient losses to the environment.
Accurately identifying individual differences in feed efficiency between growing beef cattle has historically been difficult because the interrelated aspects of physiology, genetics, environment, and economics all influence measures of feed efficiency. Individual differences in efficiency also often correspond to differences in digestion and metabolism that alter the net energy available from feed, but individual measurements of digestion and metabolism are intensive, costly and cannot be rapidly determined across large numbers of cattle.
Thus, most tools designed to identify more efficient animals have focused on estimating an animal’s genetic merit for efficiency with the goal of identifying superior animals for use in breeding.
Unfortunately, genetic merit only explains about 25% or less of the individual differences in efficiency in beef cattle and measurements of efficiency in growing cattle are only marginally related to measurements of efficiency in mature cattle. Limits in the ability to identify individual differences in efficiency from genetic tools alone is perhaps not too surprising, because nearly all genetic investigations in efficiency have been based on measures of complex biological processes (e.g., growth and intake) and have not focused on the principal factors (e.g., metabolism, energy expenditure) that contribute to efficiency.
At its core, individual differences in feed efficiency in cattle reflect the amount of feed used for production versus the total amount of feed consumed. Feed contains many different nutrients that are essential for the body, but energy or calories provided from metabolism of nutrients in feed are what predominately determine productivity in cattle that are beyond the typical age of weaning. Therefore, the calories that cattle can use for production compared to the total calories available are what largely determine differences in feed efficiency.
Cattle nutritionists typically refer to the total calories available in feed to support survival, growth, lactation and pregnancy as the feed’s “net energy”, and cattle do not use net energy equally in support of different body functions. The body processes most important to supporting survival and sustaining body tissues are collectively referred to as “maintenance” processes, and diet net energy is used first in support of these functions.
While maintenance energy requirements are critical to survival, they also represent calories from feed that were not used for production – kind of like a feed tax that cattle must pay every day to live. This tax differs between individual cattle and it’s likely that many different complex processes in the body attribute to these differences; however, because energy cannot be created nor destroyed, the amount of net energy needed to support maintenance energy requirements are identical to the amount of energy lost from the body as heat.
Additionally, since cattle are warm-blooded animals, the amount of energy lost from the body as heat is directly related to the surface area of the body.
This relationship between body surface area and maintenance energy requirements has been understood for over 100 years, but historically, measuring the surface area of an animal’s body has been difficult because cattle are not perfect geometric shapes.
Due to the difficulty in directly determining measures of surface area in cattle, an approach was adopted in the mid-twentieth century to use a mathematical manipulation of body weight as a proxy for surface area to calculate maintenance energy requirements in beef cattle.
This proxy for surface area has become more popularly known as the “metabolic body weight”, and it has several flaws that limit it from accurately determining maintenance energy requirements of cattle.
Furthermore, measures of metabolic body weight in cattle provide no further insight toward individual differences in efficiency between cattle than measures of body weight alone. Nonetheless, until recently, metabolic body weight was the most useful method available to estimate maintenance energy requirements and feed efficiency in cattle.
At the University of Missouri, we have put together a team of cattle nutritionists, geneticists and computer engineers to take a different approach to determining maintenance energy requirements in beef cattle. With advances in computer vision technology and artificial intelligence (Ai), we have been able to develop a system that can rapidly measure the surface area and volume of cattle (Figure 1).
This system allows us to determine maintenance energy requirements of beef cattle more accurately using the first principles of animal energetics. Furthermore, in using computer vision and Ai we are now able to visualize individual differences in maintenance energy requirements between cattle in real-time. This also allows us to predict the metabolic needs of different classes of cattle across various production contexts.
For example, we recently conducted an experiment where we directly measured the maintenance energy requirements of finishing steers being fed prior to harvest. In that study, we observed that estimates of maintenance energy requirements using metabolic body weight were 25% greater than direct measures of maintenance energy requirements.
Alternatively, estimates of maintenance energy requirements using computer vision and Ai provided values nearly identical to direct measures of maintenance energy requirements (7.631 vs. 7.628 Mcal/day).
These two different measures were also not statistically different from each other. Providing good evidence that visualizing differences in energy requirements with computer technology can provide more accurate values than estimations using metabolic body weight.
The use of computer vision and Ai should, therefore, enable beef producers to better determine the value of current cattle, pair cattle with different potentials for feed efficiency into different production systems to optimize economic returns, and to make breeding and culling decisions that enable more robust improvements in feed efficiency of future generations of cattle.
Ultimately, the use of this technology should increase economic returns to U.S. beef producers while decreasing the environmental footprint of beef production and increase the availability of beef to a greater number of consumers.
Brake is an assistant professor in the Animal Science Research Center at the University of Missouri.
