Amaral Thaís B, Le Cornec Alain P, Rosa Guilherme J M
Embrapa Beef Cattle, Campo Grande, MS 79106-550, Brazil.
Department of Animal and Dairy Sciences, University of Wisconsin-Madison, Madison, WI 53706, USA.
Transl Anim Sci. 2024 Aug 12;8:txae120. doi: 10.1093/tas/txae120. eCollection 2024.
The "Precoce MS" program, established by the Brazilian government in Mato Grosso do Sul in 2017, aims to encourage beef producers to harvest animals at younger ages to enhance carcass quality. About 40% of the beef produced in the state now comes from this program, which offers tax refunds ranging from 49% to 67% based on carcass classification and production system. Despite the program success, with participants delivering younger animals (with a maximum of 4 incisors), there remains significant variability in carcass quality. This paper investigates management practices and environmental factors affecting farm performance regarding carcass quality. Data from all animals harvested between the beginning of 2017 and the end of 2018 were analyzed, totaling 1,107 million animals from 1,470 farms. Farm performance was assessed based on the percentage of animals achieving grades "AAA" and "AA." Each batch of harvested cattle from each farm was categorized into two groups: high farm performance (HFP, with more than 50% of animals classified as "AAA" or "AA") and low farm performance (LFP, with less than 50% classified as such). A predictive logistic model was developed to forecast farm performance (FP) using 14 continuous and 15 discrete pre-selected variables. The most effective model, obtained through backward stepwise variable selection, had an of 0.18, accuracy of 71.5%, and AUC of 0.715. Key predictors included animal category, production system type, carcass weight, individual identification, traceability system, presence of a feed plant, location, and the Normalized Difference Vegetation Index (NDVI) from the 12-mo average before harvest. Developing predictive models of carcass quality by integrating data from commercial farms with other sources of information (animal, production system, and environment) can improve our understanding of production systems, optimize resource allocation, and advance sustainable animal production. Additionally, they offer valuable insights for designing and implementing better sectorial, social, and environmental policies by public administrations, not only in Brazil but also in other tropical and subtropical regions worldwide.
2017年，巴西政府在南马托格罗索州设立了“Precoce MS”项目，旨在鼓励牛肉生产商在牛龄较小时就进行屠宰，以提高胴体质量。该州目前约40%的牛肉产自这个项目，该项目根据胴体分级和生产系统提供49%至67%不等的退税。尽管该项目取得了成功，参与者提供的牛龄更小（最多4颗门牙），但胴体质量仍存在显著差异。本文调查了影响农场胴体质量表现的管理实践和环境因素。分析了2017年初至2018年底期间所有屠宰动物的数据，共计来自1470个农场的11.07亿头动物。根据达到“AAA”级和“AA”级的动物百分比来评估农场表现。每个农场每批屠宰的牛被分为两组：农场表现高（HFP，超过50%的动物被评为“AAA”或“AA”级）和农场表现低（LFP，低于50%被评为此类）。利用14个连续变量和15个预先选定的离散变量，开发了一个预测逻辑模型来预测农场表现（FP）。通过向后逐步变量选择获得的最有效模型，其卡方值为0.18，准确率为71.5%，曲线下面积（AUC）为0.715。关键预测因素包括动物类别、生产系统类型、胴体重量、个体识别、可追溯系统、饲料厂的存在、位置以及屠宰前12个月平均的归一化植被指数（NDVI）。通过整合商业农场的数据与其他信息来源（动物、生产系统和环境）来开发胴体质量预测模型，可以增进我们对生产系统的理解，优化资源分配，并推动可持续动物生产。此外，它们为公共管理部门设计和实施更好的部门、社会和环境政策提供了有价值的见解，不仅在巴西，在全球其他热带和亚热带地区也是如此。
