Figure 7

Multivariate statistical analysis of pairwise metabolite comparisons between the three groups. (A1–C1) Principal component analysis (PCA). If the difference between two samples is significant, the coordinate points on the score plot will be relatively far apart, and vice versa. The ellipse region represents a 95% confidence interval. (A2–C2) Partial least squares-discriminant analysis (PLS-DA). R2Y (cum) and Q2 (cum) represent the explanatory and predictive abilities of the PLS-DA model, respectively. Values closer to 1 indicate better model performance in explaining and predicting differences between the two groups. (A3–C3) Orthogonal partial least squares-discriminant analysis (OPLS-DA). The t1 component reflects the maximum inter-group differences, allowing direct differentiation of inter-group variations, while the orthogonal components reflect intra-group variations. (A4–C4) Loadings plots to identify the influence strength of metabolites on the comparative groups. The loadings range can be − 1 to 1. Loadings close to − 1 or 1 indicate a strong influence of the variable on the component, while loadings close to 0 indicate a weak influence of the variable on the component.