This protocol enables qualitative and quantitative analysis of untargeted metabolome based on FDR quality control which effectively reduces the false positives of metabolite identification. This workflow integrates XY-Meta that uses the target-decoy strategy to evaluate FDR more accurately for metabolome identification in the qualitative analysis module. This measure can effectively filter false positive results of untargeted metabolome identification, which improves the robustness of discovering biomarkers or key molecules.
We expect that researchers can understand and master the target-decoy strategy for FDR control and should try to strictly run this pipeline several times with the default parameters of the protocol. Begin by going to the GNPS database webpage and click browse datasets. Search the key word in the title column, then click the dataset ID number.
Download the dataset using FTP and save the raw data in the respective folder. To convert the format of raw data, first install the ProteoWizard software. Under the ProteoWizard installation path, type msconvert.
exe, followed by the specific parameters to convert the raw data format to mzXML format. Again, using msconvert. exe convert this data to MGF format and store them in the MGF folder.
To prepare the reference spectral library for the metabolites, go to the GNPS webpage. Search the keyword NIST, click view for the details and download the library. Save the library in the database folder.
Download the XY-Meta program. Find the parameter configuration file under the config folder and change its contents as described in the text protocol. Set the type of adducts as a list in the adduct folder.
Perform the metabolite identification and false discovery rate control using the command XY-Meta.exe. Download and install the metaX software package. Then edit the samplelist.
txt file to specify the sample and its corresponding mass spectrometry data as described in the text protocol. Use the R file provided with the text protocol to run the script for quantification of Mock and wild-type groups using the metaX software. Check the output folder in which the results of the quantitative analysis are stored such as the PCA plot.
Next, modify the parameters in the R script to annotate the peaks in qualitative and quantitative analysis using metabolite identifications in order to integrate both the results and run the R script. Box plots of quantified metabolites showed that the general distribution of healthy and disease samples was similar with low fluctuation of the mean values. Only 3.39%of the metabolites had more than 30%of the missing values.
Principle component analysis plot of the samples from both the groups showed that metaX remarkably increased the proportion of the metabolites with CV less than 0.3. A Venn diagram of differentially detected metabolites from three statistical test methods revealed 119 common metabolites. Retention time and mass by charge distribution of all annotated metabolites were plotted with a false discovery rate of less than 0.01, showing six significant and differentially detected metabolites.
It is important to limit the time for workflow testing. Remember, not to select too many samples for analysis and keep at least two samples in each group. This technique enables quality control of metabolized identification from data independent acquisition data constructing a more robust reference spectrum spectral library using the spectrum matching results based on FDR control.
