Simca p software version 11

After the data were corrected for minor variation resulting from inter-day variation in instrument tuning (Evans et al., 2009) , the missing values for a given metabolite were imputed with the observed minimum detection value, assuming they were below the limit of detection. For the convenience of data visualization, the raw area counts for each metabolite were rescaled by dividing each sample value by the median value for the specific metabolite. Furthermore, Spotfire (version 3.3.2, TIBCO Software, Palo Alto, CA, USA) was used to generate heatmaps to visualize the data (fold change and p-value versus the control). Principal component analysis (PCA) was performed for global metabolic profiles using SIMCA-P software version 11.5 (Umetrics, Umea, Sweden). Data were analyzed using JMP (SAS, http:// www.jmp.com), a commercial software package, and "R" (http://cran.r-project.org/), a freely available open-source software package. The observed relative concentrations for each metabolite were log-transformed, because, in general, the variance increased as a function of a metabolite's average response. Welch's t-tests were performed to compare data between experimental groups. The false discovery rate (FDR) method was used to account for multiple comparisons, and each FDR was estimated using q values.