Rojas de Oliveira Hinayah, Campos Gabriel S, Lazaro Sirlene F, Jamrozik Janusz, Schinckel Alan, Brito Luiz F
Department of Animal Sciences, Purdue University, West Lafayette, IN 47907.
Centre for Genetic Improvement of Livestock (CGIL), Department of Animal Biosciences, University of Guelph, Guelph, ON, N1G 2W1 Canada.
JDS Commun. 2024 Feb 1;5(3):241-246. doi: 10.3168/jdsc.2023-0460. eCollection 2024 May.
Lactation curves, which describe the production pattern of milk-related traits over time, provide insightful information about individual cow health, resilience, and milk production efficiency. Key functional traits can be derived through lactation curve modeling, such as lactation peak and persistency. Furthermore, novel traits such as resilience indicators can be derived based on the variability of the deviations of observed milk yield from the expected lactation curve fitted for each animal. Lactation curve parameters are heritable, indicating that one can modify the average lactation curve of a population through selective breeding. Various statistical methods can be used for modeling longitudinal traits. Among them, the use of random regression models enables a more flexible and robust modeling of lactation curves compared with traditional models used to evaluate accumulated milk 305-d yield, as they enable the estimation of both genetic and environmental effects affecting milk production traits over time. In this symposium review, we discuss the importance of evaluating lactation curves from a longitudinal perspective and various statistical and mathematical models used to analyze longitudinal data. We also highlighted the key factors that influence milk production over time, and the potential applications of longitudinal analyses of lactation curves in improving animal health, resilience, and milk production efficiency. Overall, analyzing the longitudinal nature of milk yield will continue to play a crucial role in improving the production efficiency and sustainability of the dairy industry, and the methods and models developed can be easily translated to other longitudinal traits.
泌乳曲线描述了与牛奶相关的性状随时间的生产模式，提供了有关个体奶牛健康、恢复力和产奶效率的深刻信息。关键功能性状可通过泌乳曲线建模得出，如泌乳峰值和持续性。此外，诸如恢复力指标等新性状可根据每头动物观察到的产奶量与拟合的预期泌乳曲线偏差的变异性得出。泌乳曲线参数是可遗传的，这表明可以通过选择性育种来改变群体的平均泌乳曲线。可以使用各种统计方法对纵向性状进行建模。其中，与用于评估305天累计产奶量的传统模型相比，随机回归模型能够对泌乳曲线进行更灵活、更稳健的建模，因为它们能够估计随时间影响产奶性状的遗传和环境效应。在本次专题综述中，我们讨论了从纵向角度评估泌乳曲线的重要性以及用于分析纵向数据的各种统计和数学模型。我们还强调了随时间影响产奶量的关键因素，以及泌乳曲线纵向分析在改善动物健康、恢复力和产奶效率方面的潜在应用。总体而言，分析产奶量的纵向特性将继续在提高乳制品行业的生产效率和可持续性方面发挥关键作用，并且所开发的方法和模型可以很容易地应用于其他纵向性状。
