An Approach for Metabonomics Data Analysis Based on Orthogonal Signal Correction Partical Least Square Discriminate Analysis

Datasets resulting from metabolomics or metabolic profiling experiments are becoming increasingly complex, which is hard to summarize and virsualize without appropriate tools. The use of chemometric tools, such as orthogonal signal correction(OSC), principal component analysis (PCA), Partial least squares to latent structure discriminant analysis (PLS-DA), make the data dimensionality reduction and interpretation much easier. Here we showed an system method based on PCA, OSC-PLS-DA for metabonomic data analysis; Furthermore, U-plot, as a visualized tool, combined with independent samples T test, were used for the biomarkers discovery. As an example, dataset from RZ water extract administrated rats urine collected by LC/MS/MS was used to demonstrate this method. As a result, U-plot based on OSC-PLS-DA was proved to be an effective, time saving tool for data interpretation and biomarkers discovery.