FIGURE 6.

Summary of natural abundance ¹³C NMR metabolomics. In (A) we show 2D STOCSY and SHY correlation maps that were produced from the same mixture of 20 common metabolites that are shown in Figure 4. We made two groups with five replicates each that had variation introduced by individual pipetting. The 2D ¹³C–¹³C STOCSY spectrum in (A) is the ¹³C statistical correlation map from all the ¹³C 1D NMR data. The 2D ¹³C–¹H SHY is the statistical correlation map between ¹³C and ¹H 1D NMR data. In (B) we present a summary of a PLS-DA of fruit flies that were either cold susceptible (blue) or cold hardy (red). The left side shows the PLS-DA using ¹³C and the right side using ¹H data. The improved ¹³C spectral resolution greatly improves the performance in multivariate analyses, such as PLS-DA. In (C) we obtained both ¹H (left) and ¹³C 1D NMR data from mice with a mutation in the mdx gene, which causes Duchenne muscular dystrophy in humans. The resonances around 3.02 ppm in the ¹H spectrum were consistent with creatine (green), but inspection of the ¹³C 1D spectra on the right shows no evidence for the corresponding ¹³C resonances, indicating that creatine is not the major component at 3.02 ppm. Thus, ¹³C data can help prevent misidentifications of metabolites. This figure was adapted from Clendinen et al. (2014), where more details about these studies can be found.