Skip to content

Auto Crude Metabolites Filter

The Auto Crude Metabolites Filter module helps researchers filter crude metabolite data using characteristic ion and characteristic NL data to build the target metabolite network.

Parameter Inputs

To use this module, users need to input the following parameters:

1. Characteristic Ion and Characteristic NL

  • Users can input either characteristic ion or characteristic NL (both cannot be empty).
  • Multiple entries should be separated by spaces.
  • The module will filter the crude metabolites based on the provided characteristic data.

2. Filter Model

This module supports two filter models: - MN Model: Only a .mgf file of the crude metabolites is required. - FBMN Model: Both a .mgf file and a corresponding .csv file (processed using tools like MZmine) need to be uploaded.

3. Match Count

  • Define the required number of matches with the characteristic data.
  • Molecules whose characteristic data matches count higher than the user-defined threshold will be filtered.

Two match modes are available: - AND Mode: Both the characteristic ion and characteristic NL must match the threshold. - OR Mode: Either the characteristic ion or characteristic NL must match the threshold.

Note:
If users only want to filter based on characteristic ion, they should still select the AND mode. Selecting OR mode would result in all molecules being filtered as every molecule satisfies characteristic NL >= 0. The same logic applies for characteristic NL filtering.

4. Minimum Normalized Intensity

  • Set a minimum normalized intensity threshold.
  • Ions with intensities below this value will be filtered out to eliminate low-intensity noise ions.

5. Tolerance

  • Define a tolerance range within which ions or neutral losses will be considered as matching with characteristic ion or characteristic NL.
  • The default value is 0.02.

6. Cosine Score

  • Set a cosine similarity threshold.
  • Compounds with cosine similarity higher than the threshold will be connected, and a molecular network preview will be generated for the filtered compounds.

Viewing and Downloading the Results

Filtered target compounds will be displayed as a molecular network in the "Preview of Filter MN" section on the right. Users can:

  • Preview the filtered compounds' m/z, retention time (rt), cosine score, and other information.
  • Download the filtered .mgf and .csv files for further analysis or upload to the GNPS platform to generate the final MZmol.

Metadata File and Classification

Upon downloading, a metadata.csv file will be generated containing a "classification" column. The numbers in this column indicate the specific classification of the molecule:

  • Number 3: Both the characteristic NL and characteristic ion match counts exceed the threshold.
  • Number 2: Only the characteristic NL match count exceeds the threshold.
  • Number 1: Only the characteristic ion match count exceeds the threshold.

Quick Tips:

  • Use the AND Mode to filter based on both characteristic ion and neutral loss.
  • Set appropriate cosine score thresholds to ensure the best network generation results.
  • Preview the molecular network to quickly identify key connections and data relationships.

This module allows for advanced filtering of crude metabolites, providing researchers with a streamlined approach to building targeted metabolite networks.