Joint genomic and metabolomic analysis of toxic dose-response experiments

Abstract

A methodology has been implemented for analyzing microarray and NMR spectral data obtained from the same set of toxic-exposure dose-response experiments. The NMR spectra additionally track the time course of exposure. Analyses consist of screening the data to eliminate variates with insignificant signal, normalization appropriate to the experimental design, principal components analysis, and nonlinear classification using a support vector machine. It is found that exposure at subtoxic levels can be detected.