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.