Impact of parity differences on residual feed intake estimation in Holstein cows

Authors: CAVANI, LIGIA - University Of Wisconsin; PARKER GADDIS, KRISTEN - Council On Dairy Cattle Breeding; Baldwin, Ransom - Randy; SANTOS, JOSE - University Of Florida; KOLTES, JAMES - Iowa State University; TEMPELMAN, ROBERT - Michigan State University; VANDEHAAR, MICHAEL - Michigan State University; CAPUTO, MALIA - University Of Wisconsin; WHITE, HEATHER - University Of Wisconsin; PENAGARICANO, FRANCISCO - University Of Wisconsin; WEIGEL, KENT - University Of Wisconsin

Submitted to: Journal of Dairy Science Communications
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/22/2022
Publication Date: 2/2/2023
Citation: Cavani, L., Parker Gaddis, K.L., Baldwin, R.L., Santos, J.E., Koltes, J.E., Tempelman, R.J., Vandehaar, M.J., Caputo, M.J., White, H.M., Penagaricano, F., Weigel, K.A. 2023. Impact of parity differences on residual feed intake estimation in Holstein cows. Journal of Dairy Science Communications. https://doi.org/10.3168/jdsc.2022-0307.
DOI: https://doi.org/10.3168/jdsc.2022-0307

Interpretive Summary: Residual feed intake (RFI) has been used as a measure of feed efficiency in farm animals. For lactating dairy cattle it is determined by an equation that tests a standard prediction against the actual nutritional intake by the cow. We tested statistical approaches to best define the impact. Data consisted of 58,954 weekly RFI records of 5,713 lactating Holstein cows collected from 2007 to 2021 in five research stations in the USA. Estimates of heritability, repeatability, and genetic correlations between weekly RFI for parities 1, 2, and 3 were obtained using bivariate repeatability animal models. The nested RFI model had the best results compared to non-nested models. We conclude that equation elements within parity when computing RFI improves model goodness-of-fit, but the impact on the estimated breeding values appears to be minimal.

Technical Abstract: Residual feed intake (RFI) has been used as a measure of feed efficiency in farm animals, and in lactating dairy cattle it is typically obtained as the difference between DMI observations and predictions from regression on known energy sinks, and effects of parity, DIM, and cohort. The impact of parity (lactation number) on the estimation of RFI is not well understood, so the objectives of this study were to (i) evaluate alternative RFI models in which the energy sinks (metabolic body weight, body weight change, and secreted milk energy) were nested or not nested within parity, and (ii) estimate variance components and genetic correlations for RFI across parities. Data consisted of 58,954 weekly RFI records of 5,713 lactating Holstein cows collected from 2007 to 2021 in five research stations in the USA. Estimates of heritability, repeatability, and genetic correlations between weekly RFI for parities 1, 2, and 3 were obtained using bivariate repeatability animal models. The nested RFI model showed better goodness-of-fit than the non-nested model, and some partial regression coefficients of DMI on energy sinks were heterogenous between lactations. However, the Spearman's rank correlation between RFI values calculated from nested and non-nested models was equal to 0.99. Heritability estimates for RFI were similar across parities, ranging from 0.25 in parity 1 to 0.21 in parity 3. Repeatability estimates for RFI across weeks within parities were high, approximately 0.60. Estimated genetic correlations for RFI between pairs of parities were 0.99 between parities 1 and 2, 0.92 between parities 2 and 3, and 0.59 between parities 1 and 3. Similarly, Spearman's rank correlations of sires' breeding values were 0.99 between parities 1 and 2, 0.98 between parities 2 and 3, and 0.78 between parities 1 and 3. We conclude that nesting energy sinks within parity when computing RFI improves model goodness-of-fit, but the impact on the estimated breeding values appears to be minimal.