GC-MS Analysis of Derivatized Metabolites

The derivatized samples were analyzed using a global unbiased mass spectrometry-based platform with GC-MS, and data normalization was performed according to a previous study (Zhang et al., 2011 (link)). The samples were randomized and the data acquisition was done in one batch to eliminate system errors. GC-MS was carried out on an Agilent 7890A/5975C GC-MS and auto-sampler unit. An HP-5MS (Agilent J&W Scientific, Folsom, CA) column with 0.25 µm thickness, 250 µm diameter, and 30 m length was used to separate derivatized metabolites. 1 µl of sample was injected in split mode in a 1:20 split ratio by the auto-sampler. The injection temperature was set at 280°C and the column oven temperature was 80°C, with helium as the carrier gas. Mass spectrometry settings were as follows: ion source temperature was 250°C, interface temperature was 280°C, and solvent cut time was 5 min. Temperature program: 5 min hold at 40°C, followed by 10°C/min ramp to reach a final temperature of 300°C held for 5 min. The scan range was 35-750 m/z.
Raw GC/MS data were converted into CDF format (NetCDF) files by Agilent GC/MS 5975 Data Analysis software and subsequently processed by the XCMS (www.bioconductor.org) using XCMS's default settings with the following changes: xcmsSet (fwhm=3, snthresh=3, mzdiff=0.5, step=0.1, steps=2, max=300), group (bw=2, minfrac=0.3, max=300). The signal integration area of each metabolite was normalized to the internal standard (ribitol) for each sample. Identification of metabolites using the automated mass spectral deconvolution and identification system (AMIDS) was searched against commercial available databases such as National Institute of Standards and Technology (NIST) and Wiley libraries. Metabolites were confirmed by comparison of mass spectra to the spectra library using a cut-off value of 70%, and by matching the experimental retention time index (RI) of each metabolite with the mass spectral and RI collection of the Golm Metabolome Database (GMD). Principle component analysis (PCA) was performed with R software (www.r-project.org). Heat map packages available in R were used to draw heat maps, and Mev (MultiExperiment Viewer) 4.8 software was used to perform a one-way ANOVA with standard Bonferroni correction. Identified metabolites were mapped onto general biochemical pathways according to annotation in KEGG. Metabolic network maps were constructed by incorporating the identified and annotated metabolites using Cytoscape 3.2.0 software (http://www.cytoscape.org/).