LC-MS and GC-MS Metabolomic Data Analyses

ProteoWizard software was used to transform the original LC-MS data to mzXML format. The data were processed by XCMS. GC-MS raw data were processed by Chroma TOF software. After peak identification, peak alignment, peak extraction, retention time (RT) correction, and peak integration, a three-dimensional data matrix was obtained. To make the metabolomics data reproducible and reliable, peaks with relative standard deviations greater than 30% in the QC samples were filtered out. The remaining peaks were identified by comparison of RT and mass to charge ratio (m/z) indexes in a library containing spectral information from the online database of HMDB⁴, Kyoto Encyclopedia of Genes and Genomes (KEGG)⁵, and the in-house library. The GC-MS data were matched with the LECO-Fiehn Rtx5 database. Peak intensity was quantified by using the area under the curve. The data matrix was further processed by removing the peaks with missing values in more than 50% of the samples and substituting the remaining missing values with half of the minimum value. Then, a new data matrix was generated by normalizing the data to the peak intensity of the internal standard.