Fig. 2

Sample classification by principal component analysis and overview of metabolomics profiling. a Principal component analysis (PCA) of all the m/z features measured in positive and negative ionization mode. The variation retained by PC1 (16.3%) is represented of the X axis and the variation retained by the PC2 (9.3%) is represented on the Y axes. Ellipses represent the 95% confidence interval for each group. b Heatmap generated with scale and centered normalized m/z abundances measured in positive and negative ionization mode. Euclideian distance between groups and Minkowski distance between metabolites was used and clustering was calculated with the Ward (Ward.D2) method for minimum variance between m/z values. c Principal component analysis (PCA) of the annotated m/z features in positive and negative ionization mode. The variation retained by PC1 (16.7%) is represented of the X axis and the variation retained by the PC2 (13.5%) is represented on the Y axis. Ellipses represent the 95% confidence interval for each group. d Heat map generated with scale and centered normalized m/z abundances of the annotated metabolites. Same methods used for Fig. 2b were applied. For statistical details see Methods section