Ren Xiaoli, Chu Chu, Bao Xiangnan, Yan Lei, Bai Xueli, Lu Haibo, Liu Changlei, Zhang Zhen, Zhang Shujun
Key Lab of Agricultural Animal Genetics, Breeding and Reproduction of Ministry of Education, Huazhong Agricultural University, Wuhan 430070, China.
Frontiers Science Center for Animal Breeding and Sustainable Production, Huazhong Agricultural University, Wuhan 430070, China.
Animals (Basel). 2025 Jul 30;15(15):2242. doi: 10.3390/ani15152242.
The somatic cell count (SCC) and differential somatic cell count (DSCC) are proxies for the udder health of dairy cattle, regarded as the criterion of mastitis identification with healthy, suspicious mastitis, mastitis, and chronic/persistent mastitis. However, SCC and DSCC are tested using flow cytometry, which is expensive and time-consuming, particularly for DSCC analysis. Mid-infrared spectroscopy (MIR) enables qualitative and quantitative analysis of milk constituents with great advantages, being cheap, non-destructive, fast, and high-throughput. The objective of this study is to develop a dairy cattle udder health status diagnostic model of MIR. Data on milk composition, SCC, DSCC, and MIR from 2288 milk samples collected in dairy farms were analyzed using the CombiFoss 7 DC instrument (FOSS, Hilleroed, Denmark). Three MIR spectral preprocessing methods, six modeling algorithms, and three different sets of MIR spectral data were employed in various combinations to develop several diagnostic models for mastitis of dairy cattle. The MIR diagnostic model of effectively identifying the healthy and mastitis cattle was developed using a spectral preprocessing method of difference (DIFF), a modeling algorithm of Random Forest (RF), and 1060 wavenumbers, abbreviated as "DIFF-RF-1060 wavenumbers", and the AUC reached 1.00 in the training set and 0.80 in the test set. The other MIR diagnostic model of effectively distinguishing mastitis and chronic/persistent mastitis cows was "DIFF-SVM-274 wavenumbers", with an AUC of 0.87 in the training set and 0.85 in the test set. For more effective use of the model on dairy farms, it is necessary and worthwhile to gather more representative and diverse samples to improve the diagnostic precision and versatility of these models.
体细胞计数（SCC）和体细胞分类计数（DSCC）是奶牛乳房健康状况的指标，被视为乳腺炎诊断的标准，可区分健康、疑似乳腺炎、乳腺炎以及慢性/持续性乳腺炎。然而，SCC和DSCC检测采用流式细胞术，成本高且耗时，尤其是DSCC分析。中红外光谱（MIR）能够对牛奶成分进行定性和定量分析，具有成本低、无损、快速和高通量等显著优势。本研究的目的是建立基于MIR的奶牛乳房健康状况诊断模型。使用CombiFoss 7 DC仪器（丹麦希勒勒德FOSS公司）对奶牛场采集的2288份牛奶样本的牛奶成分、SCC、DSCC和MIR数据进行分析。采用三种MIR光谱预处理方法、六种建模算法以及三组不同的MIR光谱数据进行各种组合，以建立多个奶牛乳腺炎诊断模型。使用差值（DIFF）光谱预处理方法、随机森林（RF）建模算法和1060波数建立了有效识别健康奶牛和患乳腺炎奶牛的MIR诊断模型，简称为“DIFF-RF-1060波数”，其在训练集中的AUC达到1.00，在测试集中达到0.80。另一个有效区分乳腺炎奶牛和慢性/持续性乳腺炎奶牛的MIR诊断模型是“DIFF-SVM-274波数”，其在训练集中的AUC为0.87，在测试集中为0.85。为了在奶牛场更有效地使用该模型，有必要且值得收集更多具有代表性和多样性的样本，以提高这些模型的诊断精度和通用性。
