Correction of ultrasound measurements of carcasses in New Zealand rabbits with different covariates using a mixed linear model

Abstract

Abstract- The objective was to evaluate the use of length and depth of the rib eye area as covariates to correct estimates of the rib eye area of the Longissimus dorsi muscle in New Zealand rabbits using mixed models. Five rabbits were used at 104 days of age, which were slaughtered after ultrasound analysis to compare with the actual measurement of the area of a section of the Longissimus dorsi muscle cut (ribeye area). The analyzes were carried out using mixed models in a randomized complete block design, with five treatments (operators), as a fixed effect, and five blocks (animals), as a random effect, considering: the lack of covariates; with a covariable rib eye length; with the covariate rib eye depth; and with both covariates in the statistical model. The inclusion of the length covariate provided greater experimental precision, demonstrated by the coefficient of variation (1.54%), in addition to the values of the adjusted or unadjusted coefficients of determination, 0.84 and 0.83, respectively. This covariate presence also provided lower values in the corrected or uncorrected Akaike and Schwarz Bayesian choice criteria. With a reduction of 126% in the Akaike information strategy and 117% in the corrected Akaike, 142% in the Bayesian, compared to the model without the covariates, in addition to less exclusion. The inclusion of the covariates length of the loin eye area measured obtained via ultrasound for the correction of the Longissimus dorsi muscle area in mixed models improved experimental precision. The effect of those evaluated on the measurement of the area of this muscle was detected only with the use of this covariate, and its use is recommended.

Keywords: ancova, carcass evaluation, rabbit farming, longissimus dorsi, meat production.