Longitudinal analysis of cumulative residual feed intake in Nellore cattle

Abstract

The objective of this work was to evaluate random regression models with linear spline functions for the estimation of genetic parameters of residual feed intake (RFI) in Nellore bovines. The RFI was estimated across five test periods (at 0–14, 0–28, 0–42, 0–56, and 070–days) in 929 Nellore bulls. The variance components were estimated, and the models were compared using the deviance information criterion. Rankings and the percentage of the top 10% animals across all periods were assessed using Spearman’s correlations. The best-fit model included linear splines and heterogeneous residual variance (four classes). Heritability estimates for RFI ranged from 0.39 to 0.55, with high genetic (0.83–0.98) and phenotypic (0.82–0.98) correlations between adjacent periods. Rankings were consistent between 56 and 70 days (Spearman’s correlations = 0.99; 93.55% overlap). The test period can be shortened to 56 days without compromising genetic parameters, optimizing feed intake evaluations. The random regression models effectively model the RFI, showing sufficient genetic variance for selection.