Efficiency can be defined as achieving the greatest output with the least input; a commonly used measure is pounds of gain per pounds of feed, abbreviated as G/F. Residual feed intake was introduced in the 1960s as an alternative to G/F. This method of determining efficiency uses a mathematical equation to estimate feed intake calculated from average bodyweight and average daily gain (ADG) in a feeding study. If cattle consume less than predicted, they are presumed to be more efficient, and intakes greater than predicted indicate lesser efficiency. The equation (E1 for purposes of this article) looks like this:

Rossow heidi
Professor / School of Veterinary Medicine / University of California – Davis
Independent Nutritionist / Le Grand, California

Dry matter intake = Intercept + a × average daily gain + b × bodyweight raised to the 3/4 power + residual

The relationships among dry matter intake (DMI), gain and bodyweight are analyzed from individual animal data; the residual is a property of an individual as well. The intercept, a and are properties of all animals in a study. There are, however, problems with interpretation of these terms. The intercept is included to satisfy certain statistical requirements and has no biological meaning. It may be zero, positive or negative. In the case that there are no cattle, gain = 0 and bodyweight = 0, if the intercept is positive, nonexistent cattle consume feed and are very inefficient. On the other hand, if negative, nonexistent cattle produce feed, the essence of efficiency. Error associated with estimated intake for an individual is represented by the residual: negative = more efficient, positive = less efficient. If the equation does not explain any change in intake as related to changes in gain and bodyweight, by definition, cattle consuming the least are the most efficient. On the other hand, if it were possible to explain all differences, then cattle are equally efficient because all residuals equal 0. 

The interpretations of a and also do not make sense. Coefficients a and do not vary, which means, according to the model, composition of gain and maintenance do not vary with different feed intakes, but composition of gain has profound effects on feed efficiency. The importance of estimating composition of gain in determining efficiency is shown in the following example:

Two pens of cattle, each weighing 800 pounds and of equal body composition (18% protein, 18% fat) gain 3.3 pounds per day; the group eating less feed (8.5% less feed consumed) has gains that contain 37% fat while the group eating more has gains containing 44% fat. Cattle are slaughtered after 180 days on feed weighing 1,400 pounds. The “more efficient” group has approximately 26% empty body fat, and the “less efficient” group 29% empty body fat. Cattle with 26% empty body fat will grade Select, and those with 29% empty body fat will grade Choice. Calculated differences in DMI are approximately 300 pounds – at current ration costs, less than $75 over the feeding period. The current Choice/Select spread is $28.89 per hundredweight. Differences in carcass value are approximately $250 per head. Since more feed is required to lay down fat compared to protein, animals that grade Choice, in this example, appear less efficient even though their carcasses have greater value.

Advertisement

To determine if residual sign (plus or minus) and magnitude are due to an improperly specified equation, we looked at seven datasets. Five of these used the GrowSafe system; in one study cattle were individually fed ad libitum, and in the last study cattle were individually fed from maintenance to ad libitum. Using the equation above, we sorted cattle in each study into five groups by residual (least to greatest). In our model (E2), feed required for gain for each group was estimated using an equation from the most recent “Nutrient Requirements of Beef Cattle,” in which both gain and bodyweight are taken into consideration. Maintenance was not calculated in the common manner as b × bodyweight raised to the ¾ power, but rather as (bodyweight)x, where the value of x varied among groups but was between 0.40 and 0.50. We evaluated a number of ways to describe maintenance, and this method gave us an answer that best described the data.

Our analysis indicated that cattle deemed to be the most efficient, based on residuals, ate less (3.4 pounds per day) and gained less fat. Our calculations show that “efficient” cattle had gains containing 40.9% fat, and the least efficient cattle had 48.9% fat in gains. The ratio (40.9/48.9 = 0.836) is very similar to that in the example found in the third paragraph of this article (0.84), in which we demonstrate the possibility that lesser feed intakes mean cattle have less with which to work. In six of the seven studies, energy gain was not determined. For the seventh study it was, and our estimates of energy gain were not different from those reported for those data in the literature. In E2, unlike the assumption in E1, as cattle eat more, they use some of that energy to lay down fat. It was also noted that variability in composition of gain was associated with poor prediction of intake in E1.

Estimates of maintenance were determined to be variable and increased as residuals increased, which differs from the static description of maintenance in E1 and in most feeding systems. Estimated maintenance requirements were 13.8 Mcal ME for efficient cattle and 16.1 Mcal ME for inefficient cattle. We have previously shown that energy and DMI are drivers of maintenance, similar to what is seen here.  

Based on a mathematical method evaluating how well models fit the data from which they were generated, our model better estimated intakes, compared to E1, for six of seven datasets. Differences in energy required for maintenance and fat gain accounted for more than 90% of differences in intakes between the “most” and “least” efficient groups. Residual ranking, used to determine efficiency, was unrelated between models; the correlation was essentially 0. If efficiency is a phenotypic property of the animal, like color, then no matter how efficiency is measured, results should be similar; in our study, they are not. Our analysis indicates that residuals, as a measure of efficiency, may be statistical artifacts of models used to estimate intake. Selection for cattle deemed efficient by E1 may produce cattle that eat less feed, but ultimately produce significantly less profit.