A detailed manuscript is under preparation. Please cite the GNPS paper in the meantime.
THIS PAGE IS NOT MAINTAINED ANYMORE - PLEASE CONSULT THE NEW GNPS DOCUMENTATION HERE.
NEW: the metadata can now be added to the Feature Based Molecular Networking workflow using that workflow experimental
- 1 Introduction
- 2 Feature Based Molecular Networking data analysis workflow (March 2018)
- 2.1 Feature Based Molecular Networking data analysis workflow: OFFLINE version
- 2.1.1 Mass spectrometry data pre-processing with MZmine2, OpenMS (in Optimus), MS-DIAL, or MetaboScape
- 2.1.2 Mass spectrometry data pre-processing with MZmine2 [Currently recommended]
- 2.1.3 Mass spectrometry data pre-processing with Optimus workflow (OpenMS) [Alternative to MZmine2]
- 2.1.4 Molecular networking and MS/MS spectral library annotation on GNPS
- 2.2 Visualize the Feature Finding Molecular Networking with Cytoscape
The Feature Finding Molecular Networking bringz together LC-MS feature detection tools (MZmine2, OpenMS, MS-DIAL, XC-MS, MetaboScape) , molecular networking (GNPS, http://gnps.ucsd.edu), and other in silico annotation tools, such as Sirius, CSI:FingerID, or Network Annotation Propagation.
The key improvements of the new Feature Finding Molecular Networking are:
(1) importing information derived from the feature detection tools into the molecular networks,
(2) discriminate isomers by retention time and remove isotopic peak,
(3) allows the annotation MS/MS of spectra with in silico tools and mapping in the molecular networks.
The new GNPS data analysis workflow is available in two versions, offline [available] and online [not available yet].
[A] Feature Based Molecular Networking as a offline version [Available] Mass spectrometry data pre-processing is performed first with on this software: MZmine2, OpenMS (in Optimus), or MS-DIAL , or MetaboScape (Bruker Daltonics), and processing the output on GNPS for molecular networking, MS/MS spectral library matching, and in silico annotation on GNPS. This version allows maximum flexibility.
[B] [Not available yet] Feature Finding Molecular Networking as an online version on GNPS web-platform [available late August 2017]. Mass spectrometry data pre-processing (MZmine2), molecular networking and MS/MS spectral library matching, and in silico annotation (Sirius), directly on GNPS web-platform are performed automatically in a single workflow. This solution is the easiest, and allows to analyze large LC-MS/MS datasets, but the user has less flexibility in the tuning of each tools.
To ensure best results, we recommend the following. First, optimize the parameters of each tools in the offline version of the software on a subset of representative samples. Then, run the online version of the workflow on the full dataset with the optimized parameters.
In summary, these tools have been adapted by providing an .MGF export module for the result of LC-MS/MS feature detection.
1) The data have to be processed as recommended by the developers.
2) The spectral data from the aligned data can be exported as .MGF, the aligned quantification table has to be exported as well.
3) The .MGF file can be used as input for any GNPS tools. The aligned quantification table can be imported into Cytoscape.
4) The metadata table groups can be generated automatically with a dedicated workflow for MZmine2, or post-processed with other solutions (Jupyter notebooks, excels).
Currently, we are recommending using the MZmine2 workflow, as it has been thoroughly tested. See the documentation below.
Download the latest MZmine2 toolbox (v2.30 minimum) at https://github.com/mzmine/mzmine2/releases
See documentation and videos here: http://mzmine.github.io/documentation.html
Start by converting your files to mzML format. See the corresponding documentation.
In MZmine2, a sequence of steps must be performed. The following batch method (.XML format) can be downloaded (temporary email firstname.lastname@example.org ) and imported into MZmine2.
The steps required for data pre-processing with MZmine2 for GNPS are shown in the screen capture below (batch method) and described after.
1. Import the files. Menu: Raw data methods / Raw data import / "Select the files"
2. Perform mass detection on MS level 1: Menu: Raw data methods / Mass detection / Set filter : MS level 1
[IMPORTANT] Set a intensity threshold at value corresponding to the triggering of the MS2 scan event.
3. Perform mass detection on MS level 2. The same masslist name can be used: Menu: Raw data methods / Mass detection / Set filter : MS level 2.
[IMPORTANT] Make sure to set an intensity threshold representative of noise level in the MS2 spectrum. Inappropriate intensity threshold could hamper the GNPS and Sirius export modules. For that reason, set it as low as possible(Example: QTOF: 100). If you have any doubt, set it to 0.
4. Build chromatogram. Menu: Raw data methods / Chromatogram builder
5. Deconvolute the chromatograms. Menu: Peak list methods / Peak detection / Chromatogram deconvolution
[IMPORTANT] tick both options "m/z range for MS2 scan pairing (Da)" and "RT range for MS2 scan pairing (min)". Define the values according to your experimental setup.
6. Deconvolute co-eluting ions "Isotopic peaks grouper module" [recommended] or Camera module. Menu: Peak list methods / Isotopes / Isotopic peaks grouper.
7. Order the peaklists. Menu: Peak list methods / Order peak lists.
8. Aligned features. Menu: Peak list methods / Alignment / Join aligner
9. Detect missing peaks. Menu: Peak list methods / Gap filling / Peak finder
10. Export the feature table containing all the peaks in .CSV format. Menu: Peak list methods / Export / Export to CSV file
[IMPORTANT] If any other filtering of the peaklist has been performed, make sure to before, reset the peak row number. Menu: Peak list methods / Filtering / Peak list row filter / Reset the peak number ID
[IMPORTANT] This feature table will not be used in further workflow. It is exported in order to do further statistical analysis.
11. Use both filters in the peaklist row filter module: 'Keep only peaks with MS2 scan (GNPS)" and "Reset the peak number ID". Menu: Peak list methods / Filtering / Peak list row filter
12. Export the feature table containing only peak with MS2 scan associated (.CSV format). Menu: Peak list methods / Export / Export to CSV file
[IMPORTANT] Include data file elements, such as "Export the peak area", or other relevant informations.
13. Export the .MGF file for GNPS. Menu: Peak list methods / Export / Export for GNPS
[IMPORTANT] Make sure to select the correct masslist (having MS2 level).
14. Export the .MGF file for Sirius. Menu: Peak list methods / Export / Export for SIRIUS
[IMPORTANT] The Sirius export module changed in MZmine2.30. Select the Merging mode: "Merge all MS/MS belonging to the same feature"
[IMPORTANT] This module will work only if isotopic grouper was used at step 6. Make sure to select the correct masslist (having MS2 level).
15. (Optional) Save your project. Menu: Project / Save project.
- Connect to GNPS http://gnps.uscd.edu and create an account.
- Upload the .MGF file exported with MZmine2 and upload the file on GNPS. See the documentation here.
- Select the uploaded .MGF file.
- Select the quantification table (feature table from MZmine2, .CSV file).
- Select the metadata table. See the documentation for the format. Adva
- Set the precursor ion mass tolerance and a fragment ion mass tolerance to "0.02" Da or equivalent to achieve appropriate spectral matching in high-resolution mass spectrometry.
- Set the Advanced Quantification Options. Used for the group mapping. Select the method for the aggregation method.
Submit the job and wait for it to be finished.
After download the Cytoscape files by clicking on "Download Cytoscape File", and explore the data.
Import the molecular network file and the annotation tables with Cytoscape 3.4 or later. Following the instructions of the https://bix-lab.ucsd.edu/display/Public/Cytoscape+3.4+Visualization+and+Analysis+Documentationdocumentation for Cytoscape 3.6.
Other annotations can be exported with MZmine2 and visualized in the molecular network, such as:
Molecular formula, adducts, or any other MZmine2 annotations can be viewed as node label: Go to Style / Node / Properties / Label / and select the column name to use.