Blaise et al. 10.1073/pnas.0707393104.

Supporting Information

Files in this Data Supplement:

SI Table 1
SI Figure 4
SI Figure 5
SI Table 2
SI Figure 6
SI Figure 7
SI Table 3




SI Figure 4

Fig. 4. Calculation of coefficients of variation. To assess intrinsic reproducibility, 15 spectra were acquired from different N2 samples. From the left to the right and from top to bottom, mean spectra (a-c), standard deviation spectra (d-f), coefficients of variation (g-i), and their distribution (j-l) are given for the whole spectrum d[-0.5; 10] (a, d, g, and j), the aliphatic region d[0.5; 4] (b, e, h, and k), and the aromatic region d[6; 9] (c, f, i, and l). The coefficients of variation show a mode around 15%, which validates the robustness and reproducibility of the overall protocol to collect 1H HRMAS-NMR spectra of different control preparations.





SI Figure 5

Fig. 5. Statistical support for the discrimination between sod1(tm776) and N2. Further study of the biological factors (genetics and age) is given with the scores (a and b), loadings (c and d), and model validation plots (e and f) related to genetics (a, c, and e) and age (b, d, and f). Spectral assignment was achieved based on literature values (1) and 2D homo- and heteronuclear HRMAS-NMR spectroscopy. Model validation for the biological factors (genetics in e and age in f) was performed by resampling the model 999 times under the null hypothesis H0.

1. Nicholson JK, Foxall PJD, Spraul M, Farrant RD, Lindon JC (1995) Anal Chem 67:793-811.





SI Figure 6

Fig. 6. Cooman's plots for age (a) and genetic effects (b). Principal component analysis (PCA) on the training set creates reference models that are displaced in the upper left and lower right of the plot. The test set is then projected on this plot, and the membership of each spectrum of the test set is then evaluated with respect to the PCA models' limits. In both cases, training sets that do not belong to the classes used in PCA to create the map are projected in the exclusion area (upper right). These SIMCA plots show the capacity of the developed protocol to identify a sample, on age (a) or genetic (b) specificity, by its comparison with previously recorded data points.





SI Figure 7

Fig. 7. Statistical support for oxidative stress pathway dissection. OPLS loadings between sod-1(tm776) and N2, ctl-1(ok1242) and N2, and sod1(tm776) and ctl-1(ok1242), respectively (a, c, and e), show variations in concentration that can be linked to metabolites. Internal validations of the above models are displayed showing a substantial decrease of performances (R2 and Q2) as genetic data are permuted (b, d, and f).