Fig. 1

Multivariate data analysis of the metabolites. a Principal Component Analysis (PCA) scores plots (left panels) and Partial Least Square-Discriminant Analysis (PLS-DA) scores plots (right panels) of ARDS patients vs healthy controls. Shaded areas are the 95% confidence regions of each group. b PCA and PLS-DA scores plots of the survivors vs the non-survivors of ARDS. Shaded areas are the 95% confidence regions of each group. c The cross validation test for PLS-DA model of the ARDS patients vs the healthy controls. d The cross validation test for PLS-DA model of the survivors vs non-survivors of ARDS. Deep blue represents the cross-validated R² (also known as Q²),pink represent the sum of squares captured by the model (R²), and light blue represents the prediction accuracy. Red star indicates the best component number in building the model. In PLS-DA of the ARDS patients vs the healthy controls analysis, R² = 0.93 and Q² = 0.86, the survivors vs the non-survivors of ARDS analysis, R² = 0.89 and Q² = 0.49, respectively