Fig. 4

Scores plots for multivariate statistical analyses of metabolites in urine. Spot urine samples were collected at 4- and 8-week post STZ injection. Panel a: Unsupervised principal components analysis (PCA) demonstrating overall structure and variability within the large LC-MS dataset for all, 4 week only, and 8 week only urine samples in either positive or negative electrospray ionization modes. Panel b: Side-by-side scores plots for the unsupervised PCA and supervised OPLS-DA multivariate analyses for each genotype in positive mode. Panel c: Side-by-side scores plots for the unsupervised PCA and supervised OPLS-DA multivariate analyses for each genotype in negative mode. These plots visualize (along with Tables 1, 2 and 3 inputs and fit statistics to aid interpretation) the STZ treatment effect in the data subset PCA and the effectiveness of OPLS-DA to uncover metabolite variables (loadings) capable of discriminating the two treatment groups