Figure 2: (a) Partial least-squares discriminant analysis (PLS-DA) scores plot (a) obtained from the mass spectrometry data of the sera from study subjects. Data allowed clear discrimination between the subjects at week 0 (open square) and at week 12 (filled square). Outlying samples of the ellipse region with the 95% confidence interval were excluded by the Hotelling’s test. , , and . The PLS-DA model was validated by a permutation test: -values and intercepts of () and (). (b) S-plot covariance  and reliability correlation from PLS-DA models (b) and loadings plots (b). The numbers for the metabolites are as given in Table 2.