Simca software package

The LC-MS raw data was analyzed by the Progqenesis QI v2.3 software (Nonlinear Dynamics, Newcastle, UK) for extracting characteristic ions of metabolites. Data processing parameters were set as following: precursor tolerance: five ppm, fragment tolerance: 10 ppm, product ion threshold: 5%. The LC-MS data was uploaded to SIMCA software package (version 14.0, Umetrics, Umeå, Sweden) for statistical analysis. The LC-MS data were obtained according to three-dimensional datasets including m/z, peak retention time (PRT) and peak intensities, and PRT-m/z pairs were used as the identifier for each ion. Metabolite identification was based on Plant Metabolome Database.^{23 (link)} Hierarchical cluster heatmap analysis (HCA), principal component analysis (PCA) and (orthogonal) partial least squares-discriminant analysis ((O)PLS-DA) were used to visualize the differential metabolites among the samples. In (O)PLS-DA analysis, variable importance in projection (VIP) and p-value can be used to select differentially expressed metabolites (DEMs) in seeds and bark, while the metabolites had the values of VIP > 1, p-value < 0.05 and fold change ≥ 2 or fold change ≤ 0.5. HCA and volcano plots were performed in the R software version 4.1.2 (www.r-project.org). Enrichment analysis was performed using MetaboAnalyst (www.Metaboanalyst.ca).

The metabolome data was analyzed using Progenesis QI software (Nonlinear Dynamics, Newcastle, United Kingdom). Simca software package (version 14.0, Umetrics, Umeå, Sweden) was used to analyze the combined positive and negative data. Human Metabolome (http://www.hmdb.ca/), METLIN (http://metlin.scripps.edu) and LIPID MAPS (http://www.lipidmaps.org/) databases were used for metabolite identification. An orthogonal partial least squares discriminant analysis (OPLS-DA) was conducted to identify potential biomarker variables. Significantly different metabolites between groups were determined by variable importance in projection (VIP) ≥ 1 and an absolute Log₂FC (foldchange) ≥ 1. A hierarchical clustering analysis was conducted with R (http://www.r-project.org/).

The metabolome of mung bean sample was extracted with a mixed solution (methanol: chloroform: water = 5: 2: 2) according to the method described in our previous study^{45 (link)}. Each group (6-HAI, 3-DAI and 6-DAI) contained 6 replicate samples. The extract was dried in a speed vacuum concentrator, and derivatized with N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA) following the protocol described before^{46 (link)}. 1 µL for each sample was injected into a DB-35MS UI capillary column (30 m × 0.25 mm, 0.25 μm, Agilent) at 280 °C in split mode (50: 1) with helium carrier gas flow set to 1 mL/min. The temperature was isothermal for 5 min at 85 °C, followed by 8 °C per min ramp to 205 °C, held at this temperature for 5 min, and then ramped at 8 °C per min to 300 °C, held for 5 min. The mass range was from m/z 60 to 1000. The temperature of transfer line and ion source were set according to the previous study^{46 (link)}. GC-MS data analysis was conducted using Chroma TOF 4.3X software of LECO Corporation (Saint Joseph, MI, USA). The metabolites were annotated based on LECO-Fiehn Rtx5 database, NIST library, and in-house database. The multivariate statistical analyses of the metabolome data were performed using SIMCA software package (V14, Umetrics AB, Umea, Sweden).

The raw data were converted to common data format (mzML) files using the software program MSconverter, and metabolomic data were obtained using the software XCMS 1.50.1 version. The positive and negative data were combined to obtain a combined data set that was imported into the SIMCA software package (version 14.0, Umetrics, Umeå, Sweden). All samples were tested to visualize the metabolic alterations by principal component analysis and (orthogonal) partial least-squares-discriminant analysis (O)PLS-DA. Variable importance in the projection (VIP) ranks the overall contribution of each variable to the OPLS-DA model, and those variables with VIP >1 were considered relevant for group discrimination. Reference material databases built by Dalian Institute of Chemical Physics, Chinese Academy of Sciences, and Dalian ChemData Solution Information Technology Co., Ltd., HMDB and METLIN, were used. In addition, metabolite pathways were searched on noncommercial databases (KEGG, http://www.genome.jp/KEGG/pathway.html).

MS raw data (total ion current, TIC) was converted into file format using ChemStation (version E.02.02.1431, Agilent Technologies Inc). ChromaTOF (version 4.34, LECO, St Joseph, MI) was used to analyze the data, and NIST and Fiehn database were used for the annotation of the metabolites. After alignment with Statistic Compare component, the ‘raw data array’ (.cvs) was obtained from raw data including peak names, retention time-m/z and peak intensities. All internal standards and pseudo positive peaks were removed. Data was transformed by log2 and then imported into SIMCA software package (14.0, Umetrics, Umeå, Sweden). Unweighted principle component analysis (PCA) and (orthogonal) partial least-squares-discriminant analysis (OPLS-DA, with sevenfold cross validation and response permutation testing, 200 times randomly permutated) were performed to visualize the metabolism difference between groups. Metabolites with variable important in projection (VIP) > 1 and p value < 0.05 by two-tailed Student’s t-test were used for identification of differential metabolites. Metabolites between groups with |fold change (FC)|≥ 1 were considered as differential metabolites. The KEGG pathways associated with the differential metabolites were identified from KEGG database (https://www.genome.jp/KEGG/pathway.html) with the threshold of p < 0.05.

Sixty milligrams of florets of PA64S (F) and PA64S (S) at stage 7 were used to dissolve the extracted lipids. A Nexera UPLC (Shimadzu, Kyoto, Japan) system fitted with a Q-Exactive quadrupole-orbitrap mass spectrometer equipped with a heated electrospray ionization (HESI) source (Thermo Fisher Scientific, Waltham, MA, USA) was used to analyze the lipidomic analyses. The acquired LCsingle bondMS raw data were analyzed by the progenesis QI software (version 2.3, Waters Corporation, Milford, MA, USA). The positive and negative data were combined to obtain combined data that were imported into the SIMCA software package (version 14.1, Umetrics, Umeå, Sweden). The VIP value of the first principal component of the OPLS-DA model was greater than 1, and the p-value of the t test was less than 0.05 as the criteria for screening differential lipids. Each sample was represented by five biological replicates.

After log2 conversion (0 was replaced by 0.000001, then converted), the values in the data matrix were imported into the SIMCA software package (version 14.0, Umetrics, Ume, Sweden). Unsupervised principal components analysis (PCA) was used to observe the overall distribution of samples and the stability of the whole analysis process. Then, supervised (orthogonal) partial least squares discriminate analysis (OPLS-DA) was used to distinguish the overall differences of metabolic profiles among groups and to find the different metabolites among groups.