Mass profiler software

The full-scan analysis was performed in triplicate using the QC samples acquired over the range of 100 to 1050 m/z in the NEG mode to create an inclusion list to further create the auto MS/MS acquisition mode. Data were extracted using batch recursive feature extraction algorithm in MassHunter Profinder B.08.00 software (Agilent) and after evaluation exported as CEF (Cluster Exchange Format) files. The features were aligned on Mass Profiler software (Agilent) using retention time (RT) tolerance of up to 0.3 min and mass tolerance of ± 15 ppm. Features with a 100% occurrence in the replicates were used to create a target MS/MS inclusion list.

Raw UPLC-MS data was first processed using Mass Profiler software (Agilent), including peak detection, peak alignment and peak integration. Subsequently, peak areas of all samples were normalized to the total area. Afterwards, a two-dimensional data matrix was generated, including retention time (RT), compound molecular weight (mass) and peak intensity. The normalized data were subjected to SIMCA-P 14.1 (Umetrics AB, Umea, Sweden) for orthogonal partial least-squares discriminant analysis (OPLS-DA) to compare metabolic differences between treatment and control. Important variables on the projection (VIP) of the OPLS-DA model were used to discover the potential variables contributing to the differentiation. Differentially expressed metabolites (DEM) were defined as VIP > 1 and p < 0.05. The METLIN web-based metabolomics database (https://metlin.scripps.edu) was used for tentative identification of significant features, including accurate mass and MS/MS spectral matching [34 (link)]. Verification of the identified metabolites was conducted using an in-house library of standards based on the accurate mass and retention time of metabolite standards provided by IROA Technologies.

Agilent Mass Profiler software was used to extract metabolic features from the LC-MS data and generate a metabolite-intensity table containing the retention time, accurate mass and intensities of all metabolites found in the samples.
Prior to any statistical analysis, data transformation and data scaling were performed on metabolic features using log transformation and auto scaling (mean-centered and divided by the standard deviation of each metabolic features) Multivariate statistical analysis, including principle component analysis (PCA) and orthogonal signal correction partial least squares discrimination analysis (OPLS-DA) was performed on SIMCA-P (version 13.0). Univariate statistical analysis, including volcano plot, fold changes, and t-test statistics were performed using MetaboAnalyst 4.0 (https://www.metaboanalyst.ca/)^{32 (link)}.
Metabolite identification was performed by matching experimental tandem MS spectra, retention time, and accurate mass of the metabolic features against in-house standard tandem MS spectra library as well as spectral databases such as METLIN and HMDB.

These data were preprocessed by Mass Profiler software (Agilent) and edited later in EXCEL2007 software. The final result is organized into a two-dimensional data matrix, including variables (retention time and mass-charge ratio), observation (sample) and peak intensity. The sample of this project obtained 6623 features in positive mode and 3581 features in the negative mode. All of the data are then normalized to the total signal integral. The edited data matrix was imported into SIMCA-P software (Umetrics AB, Umea, Sweden, version 13.0) for principal component analysis (PCA). The SAS statistical package (order no. 195557), version 9.1.3, was used for the statistical analysis. The attribute data were analysed using the χ-square test. The measurement data obtained indicated a normal distribution. Comparisons between multiple groups were analysed using analysis of variance. Metabolites with significant differences were screened by the VIP (VIP>1), p value (p < 0.01) and fold change value (FC < -3, FC > 3). Metabolic pathway analysis was performed using MetaboAnalyst software (version 3.0). The calculated p value was established on the basis of the pathway enrichment analysis whereas the pathway impact value was derived from the pathway topology analysis. Correlation analysis using was performed using the Pearson correlation coefficient with R (pheatmap package) software.

The Agilent METLIN (METLIN_AM_PCDL-N-130328.cbd) database was used to make tentative identifications from the mass lists created in Agilent Mass Profiler software. This database includes masses, chemical formulas, and structure information for various compounds. All features were searched against the METLIN database according to match m/z value. A list of identified compounds was generated.