Location: Nutrition Research
Title: An Evaluation of Residual Feed Intake Estimates Obtained with Computer Models Versus Empirical Regression Authors
Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract Only
Publication Acceptance Date: March 9, 2009
Publication Date: July 12, 2009
Citation: Williams, C.B., Ferrell, C.L., Jenkins, T.G. 2009. An Evaluation of Residual Feed Intake Estimates Obtained with Computer Models Versus Empirical Regression [abstract]. Journal of Animal Science 87 (E-Supplement 2):581. Technical Abstract: Data on individual daily feed intake, bi-weekly BW, and carcass composition were obtained on 1,212 crossbred steers, in Cycle VII of the Germplasm Evaluation Project at the U.S. Meat Animal Research Center. Within animal regressions of cumulative feed intake and BW on linear and quadratic days on feed were used to quantify average daily feed intake (ADFI) and ADG over a 120-d period. Residual feed intake (RFI) was estimated from predicted values of expected feed intake obtained by, a) empirical regression of ADFI on ADG and average BW0.75 (RFIREG), b) Cornell Value Discovery System (RFICVDS), c) National Research Council 2000 beef model (RFINRC), and d) Decision Evaluator for the Cattle Industry (DECI, RFIDECI). Observed data on growth and carcass composition were used as input to the 3 computer models. Phenotypic correlations (r = 0.95, 0.87, 0.78) for RFIREG with RFICVDS, RFINRC, and RFIDECI, respectively, suggest that RFIDECI may be a different trait from RFIREG. Additionally, RFIREG, RFICVDS, RFINRC, and RFIDECI, respectively, were correlated with ADG (r = 0.00, -0.20, -0.41, -0.48), and ADFI (r = 0.58, 0.50, 0.27, 0.09). These results show further differences between RFIDECI and RFIREG, and similarity between RFICVDS and RFIREG, with RFINRC in between. Some animals that eat very little and grow slowly were identified as efficient based on their RFIREG values, but ranged from less efficient to inefficient based on their RFICVDS, RFINRC, and RFIDECI values, respectively. These results may be due to the fact that computer models predict performance on an individual animal basis in contrast to empirical regression. Also, the formulation of maintenance in DECI results in increasing maintenance requirements with increasing BW and ADFI. Animals with very low ADFI have lower maintenance requirements with DECI and in some cases this results in expected feed intake being lower than ADFI. Finally, the results suggest that selection for RFIDECI would tend to increase ADG with no change in ADFI.