An Approach for Data Analysis of Multi-Group Metabonomics Base on Hierarchical-PCA

Datasets of metabolomics of multi-group are becoming increasingly complex, hard to summarize and visualize. Hierarchical Modeling makes the data dimensionality reduction and interpretation much easier by principal component analysis (PCA). Dose-response curve is drawed with the principal component score values. As an example, dataset from Ma Xin Shi Gan Tang (MXSGT) water extract administrated rats plasma collected by LC/MS/MS was used to demonstrate this method. As a result, Hierarchical Modeling based on PCA was proved to be an effective, time saving method for data purification.