中国荷斯坦牛乳房炎风险评估模型建立及预测分析  

Studies on Risk-assessment-model Establishment and Prediction of Mastitis in Chinese Holstein Cattle

作　　者：史良玉 李文龙 唐永杰 米思远 肖炜 刘林[3] 张毅[1] 王雅春[1] 俞英[1] SHI Liangyu;LI Wenlong;TANG Yongjie;MI Siyuan;XIAO Wei;LIU Lin;ZHANG Yi;WANG Yachun;YU Ying(College of Animal Science and Technology,China Agricultural University,Beijing 100193,China;Beijing Animal Husbandry Station,Beijing 100029,China;Beijing Dairy Center,Beijing 100085,China)

机构地区：[1]中国农业大学动物科学技术学院,北京100193 [2]北京市畜牧总站,北京100029 [3]北京奶牛中心,北京100192

出　　处：《中国畜牧杂志》2021年第3期84-90,共7页Chinese Journal of Animal Science

基　　金：国家自然科学基金国际(地区)合作项目(31961143009);北京市奶牛产业创新团队(BAIC06);北京市自然科学基金(6182021);国家奶牛产业技术体系项目(CARS-36)。

摘　　要：实验旨在探究奶牛乳房炎的风险因素,建立奶牛乳房炎风险评估模型,预测奶牛乳房炎患病风险。本研究利用北京地区1998—2016年196万余条奶牛群体改良(Dairy Herd Improvement,DHI)测定记录,将具有统计学意义的因素进行多因素Logistic回归分析,分析其中的风险因素,并建立奶牛乳房炎风险评估模型,通过大量DHI记录进行奶牛乳房炎患病风险预测研究。多因素Logistic回归分析结果发现,场规模越大,奶牛患乳房炎的概率相对越小;胎次越高,奶牛患乳房炎的概率越大;夏季奶牛患乳房炎的概率比其他季节的概率高;泌乳天数高于300 d的奶牛患乳房炎风险约为泌乳天数低于100 d奶牛的2倍。使用北京地区DHI记录进行Logistic回归分析得出,奶牛隐性乳房炎风险评估模型和奶牛临床乳房炎风险评估模型预测价值分别为0.721和0.825,认为可以应用于实际牛群。综上,本研究中确定的奶牛场规模、胎次、DHI测定季节、泌乳阶段等风险因素可用于奶牛乳房炎预测,并构建了奶牛隐性乳房炎、临床乳房炎的风险评估模型,以减少奶牛群的乳房炎发生概率。The aims of this study were to explore the risk factors of cow mastitis,establish a mastitis risk-assessment-model,and predict the risk of mastitis in Chinese Holsteins.In this study,more than 1.96 million Dairy Herd Improvement(DHI)records from Beijing dairy herds during 1998 to 2016 were used to investigate the risk factors of mastitis,to establish a risk assessment mode and to predict the risks of mastitis by using Multivariate Logistic regression analysis.Our results showed that Multivariate Logistic regression analysis disclosed that the greater the scale of the farm,the lower probability of cows mastitis,whereas the higher the parity,the higher probability mastitis.Moreover,the cows had a higher risk of mastitis in summer compared with that in other seasons.When the lactation days are higher than 300 days,the risk of mastitis in the cows is nearly twice of that in the cows with less than 100 lactation days.After Logistic regression analysis using DHI records of Holstein herds in Beijing,we found that the predicted values of the risk assessment model for subclinical mastitis and that for clinical mastitis were 0.721 and 0.825 respectively,and they were considered to be practical applications.In conclusion,the risk factors of dairy farm size,parity,DHI measurement season and lactation stage determined in this study can be used to predict mastitis in dairy cows.The risk-assessment-models of subclinical-and clinical-mastitis constructed in the study can reduce the incidences of mastitis in Holstein herds.

关 键 词：荷斯坦牛 Logistic回归分析 DHI 乳房炎风险预测 风险评估模型 

分 类 号：S823.2[农业科学—畜牧学]