Metabolomics data analysis
To visualize metabolic differences between the healthy and clinical endometritis groups, unsupervised principal component analysis (PCA) was performed based on the metabolome generated in the positive (Figure 3) and negative ion chromatograms (Figure 4). PCA can be used for the preliminary analysis of samples. Although there were differences among the various components in the PCA score plots, they were difficult to distinguish. Data analysis showed that the individual differences among dairy cows in the same group and other factors produced large intragroup variations. OPLS-DA is a supervised discriminant analysis statistical method for reducing such intragroup differences, enlarging intergroup differences, eliminating the influence of irrelevant factors on experimental data, and effectively predicting different samples. The data were processed by supervised OPLS-DA to identify differential metabolites. The parameters R2Y and Q2 were above 0.80 in the two groups, indicating that the model had reasonable interpretation and prediction ability, and the clustering separation effect of samples was noticeable.