Metabonomic Characterization of Genetic Variations in Toxicological and Metabolic Responses Using Probabilistic Neural Networks

摘要: Current emphasis on efficient screening of novel therapeutic agents in toxicological studies has resulted the evaluation analytical technologies, including genomic (transcriptomic) and proteomic approaches. We have shown that high-resolution 1H NMR spectroscopy biofluids tissues coupled with appropriate chemometric analysis can also provide complementary data for use vivo drugs. Metabonomics concerns quantitative dynamic multiparametric metabolic response living systems to pathophysiological stimuli or genetic modification [Nicholson, J. K., Lindon, C., Holmes, E. (1999) Xenobiotica 11, 1181-1189]. In this study, we used characterize time-related changes urinary metabolite profiles laboratory rats treated 13 model toxins drugs which predominantly target liver kidney. These spectra were data-reduced subsequently analyzed using a probabilistic neural network (PNN) approach. The methods encompassed database 1310 samples, 583 comprised training set network, remaining 727 (independent cases) employed as test validation. Using these techniques, classes toxicity, together variations associated strain, distinguishable >90%. Analysis spectral by multilayer perceptron networks principal components gave similar but less accurate classification than PNN analysis. This study highlighted value developing NMR-based metabonomic models prediction xenobiotic-induced toxicity experimental animals indicates possible future uses accelerated drug discovery programs. Furthermore, sensitivity tool strain differences may prove be useful investigating variation responses assessing validity specific animal models.