Substantial equivalence is a key concept in the evaluation of unintended and potentially harmful metabolic impact consequent to a genetic modification of food. The application of unsupervised multivariate data analysis to the metabolic profiles is expected to improve the effectiveness of such evaluation. The present study uses NMR spectra of hydroalcoholic extracts, as holistic representations of the metabolic profiles of grapes, to evaluate the effect of the insertion of one or three copies of the DefH9-iaaM construct in plants of Silcora and Thompson Seedless cultivars. The comparison of the metabolic profiles of transgenic derivatives with respect to their corresponding natural lines pointed out that the overall metabolic changes occur in the same direction, independent of the host genotype, although the two cultivars are modified to different extents. A higher number of copies not only produces a larger effect but also modifies the whole pattern of perturbed metabolites.