A multiway approach to analyze metabonomic data: a study of maize seeds development.

The aim of this research was to show that the application of multiway partial least square-discriminant analysis to nuclear magnetic resonance spectra is a valuable tool to analyze metabonomic data of transgenic maize. We evaluated the effects, on the development of seeds, of the introduction of the antisense-mediated downregulation and overexpression of the Rpd3 gene (ZmRpd3) in the genome of a maize inbred line, we identified the metabolites involved in the differentiation between classes of samples, directly integrating the evolution of each metabolic perturbation over time in the model. Major differences were found at the beginning of development, confirming the results obtained by transcript analysis: ZmRpd3 transcripts and proteins accumulate during the initial stage of development, suggesting a role for this gene in cell cycle control.