Heterologous Expression of Anaerobic Gut Fungal Polyketides and Nonribosomal Peptides in Model Fungal Hosts

Background

In silico genome mining tools predict many biosynthetic gene clusters for secondary metabolites from anaerobic gut fungi (phylum Neocallimastigomycota) (Swift et al. 2021). In this research project, we sought to learn more about the products of these gene clusters via the technique of heterologous expression, wherein we performed the following:

1. We inserted predicted biosynthetic genes into vectors. The predicted secondary metabolite classes were non-ribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs).
2. We transformed the vectors into model host microbes - in this project, Saccharomyces cerevisiae and Aspergillus nidulans
3. We cultured the transformed host microbes under expression conditions
4. We harvested culture samples (cell pellets) for data analysis and screening.

For this project, we implemented an untargeted metabolomics screen with LC-MS/MS (liquid chromatography with tandem mass spectrometry) to determine which expression groups possessed standout metabolites (by abundance and t-test statistics) relative to the negative control: the model host microbe expressing the corresponding empty vector (the transformed vector with no inserted gene to express).

We used two approaches to analyze the untargeted LC-MS/MS datasets: (1) Feature-based molecular networking (FBMN) (Scripts 1 - 4) (see Figure 1 below) (Nothias et al. 2020) and (2) Classical Molecular Networking (CMN) (Script 5) (Wang et al. 2017). Please refer to documentation on the GNPS (Global Natural Product Social Molecular Networking) web platform for a description of the FBMN vs. CMN approaches. Broadly, FBMN and CMN use different methods to distinguish metabolite signals in the data, with FBMN determining distinct peaks (features) for metabolites and CMN only considering MS/MS spectra. For this project and set of scripts, the CMN approach is more lenient for determining if a metabolite signal is real. For the FBMN approach, we utilized multiple existing tools to perform data processing and analysis: MZmine3 (Schmid et al. 2023), MetaboAnalyst (Xia et al. 2011), GNPS (Wang et al. 2017; Nothias et al. 2020), and the SIRIUS suite of tools (Dührkop et al. 2019), which includes ZODIAC (Ludwig et al. 2020), CSI:FingerID (Dührkop et al. 2015), and CANOPUS (Djoumbou Feunang et al. 2016; Dührkop et al. 2021; Kim et al. 2021; Hoffmann et al. 2022). For the CMN approach, we utilized GNPS and SIRIUS.

Overview Figure. Heterologous expression was implemented to study anaerobic gut fungal secondary metabolite potential. (A) Predicted NRPS and PKS core biosynthetic genes were mined from anaerobic gut fungi, inserted into expression plasmids, transformed into S. cerevisiae and A. nidulans, expressed, and screened for expected expression products (RNA/protein/metabolites). (B) The LC-MS/MS data analysis workflow with feature-based molecular networking (FBMN) systematically processed the data and compared expression groups against their corresponding empty vector control. Follow-up data analysis was performed with a classical molecular networking workflow with relatively more lenient filtering cutoffs. Figure was created with BioRender.com.

While screening heterologous expression is a main use case for this workflow, other untargeted LC-MS/MS data sources with 2 sample groups (Experimental vs. Control) can be similarly processed and analyzed. Overall, datasets run through this workflow include the following experimental-control pairings:

Experimental (EXP) Group	Control (CTRL) Group for Comparison
Heterologous Expression	Empty Vector Negative Control
Cultured Anaerobic Gut Fungi	Media Negative Control
Cultured Anaerobic Gut Fungi Spiked with Epigenetic Elicitors	Cultured Anaerobic Gut Fungi Spiked with Blank Solvent

Associated Publication

[to be filled in after publication]

Data Availability

Raw .mzML data files will be made publicly available on the MASSIVE data repository once the corresponding publication is published. (to-do: update after publication)

Script Summaries: Feature-based Molecular Networking

Script_1_MZmine3_multi-job_workflow.py

Script 1 is a Python script that automates MZmine3 preprocessing of LC-MS/MS data. It creates metadata files, modifies XML parameters (based on your manually curated parameters template file), executes MZmine3 via command line, and prepares inputs to MetaboAnalyst, GNPS, and SIRIUS analysis. The script also handles FTP upload of processed data to GNPS servers.

Script_2_MetaboAnalyst_multi-job_workflow.r

Script 2 is an R script that performs statistical analysis of metabolomics data using MetaboAnalystR. The script performs data normalization, fold-change analysis, t-tests, and principal component analysis (PCA). The script generates visualizations and statistical outputs for metabolite abundance comparisons.

Script_3_Cytoscape_networking_multi-job_workflow.py

Script 3 is a Python script that formats and filters molecular networks in Cytoscape (Shannon et al. 2003). It processes GNPS networking results, incorporates MetaboAnalyst statistical data, and applies custom visual styles. The script includes filtering options to highlight metabolites of interest and generates detailed node attribute tables.

Script_4_Volcano_Plots.py

Script 4 is a Python script that generates stylized volcano plots using MetaboAnalyst data (from Script 2).

Script Summary: Classical Molecular Networking

Script_5_Groupings_GNPS_Analysis.py

Script 5 is a Python script that formats and filters classical molecular networks from GNPS. This script enables comparison of multiple experimental (EXP) vs. control (CTRL) pairs to define metabolite features of interest for further investigation (ie: SIRIUS chemical predictions).

Script Details: Feature-based Molecular Networking (FBMN) Workflow

FBMN Part 1:

Before running Script 1:

• Organize data files:

  • (1) Manual: convert data files to .mzML format
  • (2) Manual (optional): if necessary, consider writing custom code to rename filenames from long, original names to shorter, descriptive names
  • (3) Manual: create "data" folder and populate with .mzML data files
    • Example: {directory}"LCMSMS_analysis_pipeline\data{data files folder}{data files}"
    • Note: create a separate {data files folder} sub-folder for EXP with base job name and CTRL samples
    • Note: to run the workflow with the data from the heterologous expression of gut fungal secondary metabolites, see the MASSIVE data repository once the data is public (to-do: update once published)

• Setup GNPS and WinSCP accounts to run jobs on GNPS:

  • (4) Manual (one-time): create GNPS account and WinSCP connect to massive.ucsd.edu FTP (see GNPS documentation, use same username and password as GNPS account). Consider creating a .env file in the working directory to hold the email, username, and password information (see Script 1 code for context)

• Setup metadata excel sheet with job information:

  • (5) Manual: fill out a main metadata excel sheet with running account of job information for all jobs, as well as info for all changeable values (see format of "Overall_Running_Metadata_for_All_LCMSMS_Jobs.xlsx" in the input folder)
  • Excel Columns: Job Name, Control Folder, Ionization, RT minimum cutoff, ABMBA_Feature_Name_from_Script_1, MZmine3 batch template
  • Note: determine the RT minimum cutoff by viewing the chromatograms in MZmine and qualitatively approximate the metabolites that quickly come off the column. Remove these poorly separated metabolites to improve downstream data handling and statistical analysis.

• Determine MZmine3 parameters:

  • (6) Manual: determine parameters for MZmine3 tool to run jobs and prepare a .xml parameters file template. If necessary, manually create an MZmine3 job for the data files to determine pre-processing parameters. Alternatively, use previously determined parameters.

→ Run Script 1

Script 1 features:

• Creates GNPS and MetaboAnalyst Metadata .tsv files
• Edits the template .xml parameters file for MZmine3, using information in the overall metadata excel
• Runs MZmine3 in commandline using the edited .xml parameters file
• Organizes MZmine3 output files for GNPS input
• Uploads GNPS_input_for_job_name folder to GNPS via FTP
• Uses the MZmine3 output for GNPS input to generate the MetaboAnalyst input

FBMN Part 2:

Suggested: use RStudio to run Script 2.

→ Run Script 2

Script 2 features:

• Runs the MetaboAnalyst tool to generate statistics, including log2 fold-change and raw p-values. See the MetaboAnalyst GitHub repository and MetaboAnalyst tutorial for more information.
• Creates empty GNPS_output job sub-folders for you to later manually populate with GNPS output files (see next Part).

After running Script 2:

• (1) Manual: check correctness of statistical treatments. For example, ensure appropriate feature and sample normalization (expect Gaussian curve) by going to temp folder --> job folder --> MetaboAnalystR Output --> normalization .png outputs.

FBMN Part 3:

Before running Script 3:

• (1) Manual: run the FBMN GNPS job. Once the job is complete, use the "Download Cytoscape Data" link to download the zipped GNPS outputs (see the image below). Organize these outputs in job_name sub-folders (created in Script 2) in a folder "GNPS_output". Manually unzip folders contents for GNPS outputs. These files will include the .graphml molecular networking file.
• (2) Manual (one-time): determine style .xml settings for Cytoscape networks
• (3) Manual: have the Cytoscape program open in order to run Script 3

→ Run Script 3

Script 3 features:

• Adjusts GNPS-generated molecular networks in Cytoscape using Python (py4cytoscape)
• Imports node table with additional data, creates pie charts (Cyan=EXP, Crimson=CTRL), adjusts style (ie: node borders are log-scale color-coded based on EXP average raw peak area), and labels compound names (see images below)
• Aligns data analysis results from MZmine3, MetaboAnalyst, and GNPS
• Filters metabolite features in the Cytoscape network based on possible metabolites of interest (ie: highly detected in EXP samples and not in CTRL samples)
• The following filter cutoffs are implemented for each feature, and cutoffs can be adjusted in the script:
  • (i) MetaboAnalyst-generated log2 fold-change for EXP vs. CTRL > 2.
  • (ii) MetaboAnalyst-generated raw p-value < 0.05.
  • (iii) CTRL average raw peak area < 10⁵
  • (iv) EXP average raw peak area > 10⁶
  • Note that these filtering cutoffs are suggestions for narrowing down standout metabolites and are not quantitative. For follow-up analysis, titers for metabolites of interest can be determined.
• If desired, searches data for specific compounds based on m/z and RT (with set +/- cutoffs), such as the standard ABMBA and the possible anaerobic gut fungal metabolite baumin (Swift et al. 2021).
• Generates filtered Cytoscape networks and formatted excels of features.
  • Output Excel Tabs:
    • "All Peaks Simple": unfiltered list of all peaks, with column selection simplified
    • "All Peaks": unfiltered list of all peaks
    • "Filtered Peaks of Interest": filtered peaks, using MetaboAnalyst statistics
    • "Upreg Likely Host Metabolites": filtered peaks for metabolites likely present in the control samples but detected more in the experimental samples
    • "All Cmpd Matches": all detected metabolites with MS/MS match to GNPS reference databases
    • "Cmpd Matches No Sus": all detected metabolites with MS/MS match to GNPS reference databases, with "Suspect related to..." compound annotations removed.
    • "ABMBA Standard": specifies which feature likely corresponds with the ABMBA standard added to samples
    • "Putative Baumin": specifies which feature likely corresponds to the possible anaerobic gut fungal metabolite baumin (Swift et al. 2021).
    • "Filter Parameters": a record of the specific filtering parameters used in Script 3 to generate the output excel

FBMN Part 4:

→ Run Script 4

Script 4 features:

• Generates formatted volcano plots to visualize metabolite features that are more detected in EXP or CTRL (see example below for Y-p9 positive ionization). Features that are more significantly detected in EXP are potential expression products for heterologous expression studies, and ideally these features are not detected in any CTRL samples (assuming the host cannot naturally produce the metabolite).

FBMN Part 5:

• Consider running the SIRIUS suite of tools using the MZmine3 .mgf file output for SIRIUS (.mgf file located in in temp folder, job sub-folder).
• SIRIUS predicts chemical formula, class, and structure from the MS/MS data.
• The MZmine3 feature IDs are consistent with feature IDs in the input for SIRIUS. In this way, you can generate and align SIRIUS predictions to filtered features of interest from Script 3.

Script Details: Classical Molecular Networking (CMN) Workflow

CMN Part 1:

Before running GNPS CMN Job:

• Note, these steps are the same as FBMN part 1 (organize data files before running Script 1)
• Organize data files:
  • (1) Manual: convert data files to .mzML format
  • (2) Manual (optional): if necessary, consider writing custom code to rename filenames from long, original names to shorter, descriptive names
  • (3) Manual: create "data" folder and populate with .mzML data files
    • Example:
    • "C:\Users\lazab\Desktop\python_scripts\workspace\LCMSMS_analysis_pipeline\data{data files folder}{data files}"
    • Note: create a separate {data files folder} sub-folder for EXP with base job name and CTRL samples
    • Note: to run the workflow with the data from the heterologous expression of gut fungal secondary metabolites, see the MASSIVE data repository once the data is public (to-do: update once published)
  • (4) Either (i) run Script 1 for each EXP vs. CTRL in the grouping or (ii) generate job sub-folders in the temp folder and create a metadata file {job_name + "_metadata.tsv} (columns "Filename" and "Class"). This setup is required for running Script 5, because it informs the script which CTRL set of .mzML files correspond to which EXP set of .mzML files.
    • Example metadata .tsv file: "temp" folder in directory -> sub-folder of G1 name, where G1 is an experimental group (ie: "Yeast_HE_p9_pos_20240910") -> job_name + "_metadata.tsv":

Filename	Class
2024_POS_Y-EV_1_CTRL_Run88.mzML	CTRL
2024_POS_Y-EV_2_CTRL_Run117.mzML	CTRL
2024_POS_Y-EV_3_CTRL_Run138.mzML	CTRL
2024_POS_Y-p9_1_Run100.mzML	EXP
2024_POS_Y-p9_2_Run120.mzML	EXP
2024_POS_Y-p9_3_Run155.mzML	EXP

CMN Part 2:

Run GNPS CMN Job

• (1) Manual: run the GNPS CMN job ("METABOLOMICS-SNETS-V2" workflow). Follow the GNPS documentation to upload the .mzML files and perform CMN.
• (2) Manual (one-time): first, ensure there is a "GNPS_output" folder in the directory. Then, create a "Groupings_Analysis_Folders" folder within the "GNPS_output" folder.
• (3) Manual: For each job, create a job sub-folder in the "Groupings_Analysis_Folders" folder (naming convention: "Grouping_" + job_name + "POS" OR "Grouping" + job_name + "_NEG").
• (4) Manual: Once the GNPS CMN job is complete, populate the corresponding job folders (in the "Groupings_Analysis_Folders" folder) with output files. There are two job links (see below) to use: (i) "Download Bucket Table" to get raw spectral count data values (.tsv file) and (ii) "Download Clustered Spectra as MGF" to acquire the .mgf file that can be import to SIRIUS for compound predictions. From the downloaded outputs, there will also be a folder with the .graphml molecular networking file.

CMN Part 3:

Before running Script 5:

• Setup metadata excel sheet with job information:

  • (5) Manual: fill out a main metadata excel sheet with running account of job information for all jobs, as well as info for all changeable values (see format of "Script_5_Groupings_Metadata.xlsx" in the input folder, note this is a different metadata excel from the FBMN workflow). Excel Columns: Job Name, G1, G2, G3, G4, G5, G6, G1_temp_folder, G2_temp_folder, G3_temp_folder, G4_temp_folder, G5_temp_folder, G6_temp_folder, Ionization, Cytoscape_Format_Template_File
    • G# = Group # (can be EXP or CTRL)
    • G#_temp_folder = indicate the name of the job sub-folder in the temp folder.
  • Note: I created different Cytoscape style files to format networks differently depending on the EXP vs. CTRL comparisons of interest. This information is relevant for how the script formats filtered molecular networks and visualize pie charts for relative spectral counts.

• Example 1 for Cytoscape .xml Style File:

  • "styles_7_groupings_emphasis.xml":
    • G1 = EXP1
    • G2 = CTRL1
    • G3 = EXP2
    • G4 = CTRL2
    • G5 = EXP3
    • G6 = CTRL 3
  • ie: to compare 3 different heterologous expression samples to 3 different corresponding empty vector controls

• Example 2 for Cytoscape .xml Style File:

  • "styles_7_groupings_v2.xml":
    • G1 = EXP1
    • G2 = EXP2
    • G3 = EXP3
    • G4 = CTRL (same control fro EXP1-3)
  • ie: to compare 3 gut fungal strain samples against 1 media sample

• Example 3 for Cytoscape .xml Style File:

  • "styles_7_groupings_v3.xml":
    • G1 = EXP1
    • G2 = EXP2
    • G3 = EXP3
    • G4 = CTRL for G1 and G2
    • G5 = CTRL for G3
  • ie: G1 and G2 have the same control, but G3 has a different control.

→ Run Script 5

Script 5 features:

• Takes GNPS CMN outputs to format and filter molecular networks in Cytoscape, based on EXP and CTRL groups of interest to compare (termed "groupings").
• Create node pie charts (see below) to visualize relative spectral abundances (sum precursor abundances from bucket tables) for EXP and CTRL groups to compare. Color-group pairings can be viewed in Cytoscape → select the pie chart symbol of "Image/Chart1" in the left "Node" style panel → see the "Selected Columns" which correspond to the numbered colors under "> Customize"
• Filters metabolite features based on detection in EXP(s) and lack of detection in CTRL(s). The default cutoff is > 10⁶ average sum precursor abundance in EXP and < 10⁶ average sum precursor abundance in the corresponding CTRL.
• To generate filtered molecular networks, the script combinatorially compares EXP vs CTRL filter criteria. For example, for 3 EXP vs. CTRL pairings, the script will generate 7 networks based on features that pass the 3 criteria (EXP1-EXP2-EXP3, EXP1-EXP2, EXP1-EXP3, EXP2-EXP3, EXP1, EXP2, EXP3).

CMN Part 4:

• Consider running the SIRIUS suite of tools using the GNPS output .mgf file, in order to further inspect filtered metabolites of interest.

Installation: Dependencies

Script 1: MZmine3 Multi-job Workflow

• Python 3.8+
• Required Python modules/packages:
  • pandas
  • xml.etree.ElementTree
  • subprocess
  • shutil
  • ftplib
  • python-dotenv
• MZmine3 installation
• GNPS account with FTP access

Script 2: MetaboAnalyst Multi-job Workflow

• R 4.0+
• Required R modules/packages:
  • MetaboAnalystR
  • impute
  • pcaMethods
  • globaltest
  • GlobalAncova
  • Rgraphviz
  • preprocessCore
  • genefilter
  • sva
  • limma
  • KEGGgraph
  • siggenes
  • BiocParallel
  • MSnbase
  • multtest
  • RBGL
  • edgeR
  • fgsea
  • devtools
  • crmn
  • httr
  • qs
  • readxl (for reading Excel metadata)
  • ggrepel (for plot labeling)
  • ellipse (for PCA plots)
  • vegan (for statistical analysis)

Script 3: Cytoscape Networking Multi-job Workflow

• Python 3.8+
• Required Python modules/packages:
  • pandas
  • py4cytoscape
  • numpy
• Cytoscape 3.9+ installation
• Running Cytoscape instance required during script execution

Script 4

• Python 3.8+
• Required Python modules/packages:
  • pandas
  • numpy
  • matplotlib
  • adjustText

Script 5

• Python 3.8+
• Required Python modules/packages:
  • pandas
  • numpy
  • matplotlib
  • py4cytoscape
• Cytoscape 3.9+ installation
• Running Cytoscape instance required during script execution

Tutorial

For initial testing and troubleshooting of the scripts and installed dependencies:

• Consider using the "data_tutorial" folder and "Tutorial" tabs in the metadata Excel files

• Tutorial sample comparisons:

  • For FBMN, S. cerevisiae with p9 plasmid vs. empty vector control
  • For FBMN, A. nidulans with TJGIp11 plasmid vs. empty vector control
  • For CMN (Script 5), comparison of yeast with p9 vs yeast empty vector vs. A. nidulans with TJGIp11 vs A. nidulans empty vector. p9 and TJGIp11 are vectors containing a predicted non-ribosomal peptide synthetase gene from the anaerobic gut fungal strain Anaeromyces robustus S4. Note: The metadata Excel files support multi-job analysis, with each row representing a separate experimental-control pairing to process.

• Additionally, it is useful to use the positive control sets for troubleshooting. For yeast samples, expression of 6-methylsalicylic acid synthase (6-MSAS) was used as a positive control. For A. nidulans samples, expression of PKSs for sphaerophorolcarboxylic acid (SA) and olivetolic acid (OA) was used as a positive control.

• Summary of positive control data folders from MASSIVE:

  • Yeast 6-MSAS EXP, positive ionization = "Yeast_HE_6MSAS_pos_20240910"
  • Yeast 6-MSAS CTRL, positive ionization = "Yeast_HE_EV_CTRL_pos_20240910"
  • Yeast 6-MSAS EXP, negative ionization = "Yeast_HE_6MSAS_neg_20240910"
  • Yeast 6-MSAS CTRL, negative ionization = "Yeast_HE_EV_CTRL_neg_20240910"
  • A. nidulans EXP, positive ionization = "Anid_HE_SA_OA_pos_20240910"
  • A. nidulans CTRL, positive ionization = "Anid_HE_EV_SA_OA_CTRL_pos_20240910"
  • A. nidulans EXP, negative ionization = "Anid_HE_SA_OA_neg_20240910"
  • A. nidulans CTRL, negative ionization = "Anid_HE_EV_SA_OA_CTRL_neg_20240910"

• Please see below images for summaries of expected results from the FBMN analysis workflow for the yeast and A. nidulans positive control studies:

Yeast 6-MSAS Positive Control Figure. The untargeted metabolomics screen highlighted expected polyketide 6-methylsalicylic acid (6-MSA) production in a positive control for S. cerevisiae heterologous expression. Positive and negative ionization modes were processed and analyzed separately. Negative ionization mode was more appropriate for detecting 6-MSA, due to its acidity. The following analyses highlighted feature 204 as detected in Y-POS-CTRL and not detected in the empty vector control (Y-EV): (A) Volcano plots generated by MetaboAnalystR fold-change analysis and raw p-values, (B) PLS-DA loadings, and (C) filtered feature nodes from GNPS analysis (see Materials and Methods). The SIRIUS tool predicted the correct chemical formula, C8H8O3. (D) Targeted metabolomics with 6-MSA standard also validated expected detection of 6-MSA in Y-POS-CTRL and not in Y-EV. FPS-ISTD=mix of internal standards interspersed throughout the run, control for inter-sample contamination; ExCtrl-Y=extraction control for all S. cerevisiae samples; Y-Media=SD-Ura media; Y-EV= S. cerevisiae empty vector negative control. Figure was created with BioRender.com.

Aspergillus SA/OA Positive Control Figure. The untargeted metabolomics screen highlighted expected polyketide sphaerophorolcarboxylic acid (SA) and olivetolic (OA) production in a semi-positive control (different plasmid design from the plasmids with gut fungal genes) for A. nidulans heterologous expression. Positive and negative ionization modes were processed and analyzed separately. The following analyses highlighted a molecular family largely detected in the Anid-POS-CTRL sample group and not detected in the empty vector control (Anid-EV-SA-OA): (A) Volcano plots generated by MetaboAnalystR fold-change analysis and raw p-values, (B) PLS-DA loadings, and (C) filtered feature nodes from GNPS (p-value < 0.05, log2 fold-change > 2, average raw peak area for negative control < 1E5, and average peak area for expression group > 1E5) analysis. The two highest contributing features on the PLS-DA loadings plot, features 3399 (253.1434 m/z, RT 5.4339 min) and 2699 (225.1121 m/z, RT 4.7146 min), were putatively identified as features for SA and OA, respectively. For SA (C14H20O4) and OA (C12H16O4), the SIRIUS tool predicted the correct chemical formula, class, and structure. SA was the third best structure match for feature 3399, and OA was the second-best structure for feature 2699. (D) Targeted metabolomics with SA and OA standards also validated expected detection of SA and OA in Anid-POS-CTRL and not in Anid-EV-SA-OA. Network edge thickness between nodes denotes MS/MS similarity score. FPS-ISTD=mix of internal standards interspersed throughout the run, control for inter-sample contamination; ExCtrl-Anid=extraction control for all A. nidulans samples; Anid-Media= GMM. Figure was created with BioRender.com.

Files for Comparison

• The described workflows were run for the corresponding publication, and the results are shared in the following folders:
  • "temp_for_comparison" = temp folder
  • "GNPS_output_for_comparison" = GNPS_output folder
  • "output_for_comparison_tutorial" = output folder for only the tutorial datasets (Y-p9 and TJGIp11) - for additional output files to compare to, please contact the primary author (butkovichlaza@gmail.com).
• See a description of previous jobs run in the input folder → "Overall_Running_Metadata_for_All_LCMSMS_Jobs.xlsx" excel tab "All" and "Script_5_Groupings_Metadata.xlsx" excel tab "All"

Support

For support with using these scripts, please contact butkovichlaza@gmail.com.

Authors and Acknowledgements

Primary author: Lazarina Butkovich (University of California, Santa Barbara)

Thank you to Fred Krauss for feedback and assistance in writing and formatting these scripts.

References

Djoumbou Feunang, Y.; Eisner, R.; Knox, C.; Chepelev, L.; Hastings, J.; Owen, G.; Fahy, E.; Steinbeck, C.; Subramanian, S.; Bolton, E.; Greiner, R.; Wishart, D. S. ClassyFire: Automated Chemical Classification with a Comprehensive, Computable Taxonomy. J Cheminform 2016, 8 (1), 1–20. https://doi.org/10.1186/s13321-016-0174-y.

Dührkop, K.; Shen, H.; Meusel, M.; Rousu, J.; Böcker, S. Searching Molecular Structure Databases with Tandem Mass Spectra Using CSI:FingerID. Proc Natl Acad Sci U S A 2015, 112 (41), 12580–12585. https://doi.org/10.1073/pnas.1509788112.

Dührkop, K.; Fleischauer, M.; Ludwig, M.; Aksenov, A. A.; Melnik, A. V.; Meusel, M.; Dorrestein, P. C.; Rousu, J.; Böcker, S. SIRIUS 4: A Rapid Tool for Turning Tandem Mass Spectra into Metabolite Structure Information. Nat Methods 2019, 16 (4), 299–302. https://doi.org/10.1038/s41592-019-0344-8.

Dührkop, K.; Nothias, L. F.; Fleischauer, M.; Reher, R.; Ludwig, M.; Hoffmann, M. A.; Petras, D.; Gerwick, W. H.; Rousu, J.; Dorrestein, P. C.; Böcker, S. Systematic Classification of Unknown Metabolites Using High-Resolution Fragmentation Mass Spectra. Nat Biotechnol 2021, 39 (4), 462–471. https://doi.org/10.1038/s41587-020-0740-8.

Hoffmann, M. A.; Nothias, L. F.; Ludwig, M.; Fleischauer, M.; Gentry, E. C.; Witting, M.; Dorrestein, P. C.; Dührkop, K.; Böcker, S. High-Confidence Structural Annotation of Metabolites Absent from Spectral Libraries. Nat Biotechnol 2022, 40 (3), 411–421. https://doi.org/10.1038/s41587-021-01045-9.

Kim, H. W.; Wang, M.; Leber, C. A.; Nothias, L. F.; Reher, R.; Kang, K. Bin; Van Der Hooft, J. J. J.; Dorrestein, P. C.; Gerwick, W. H.; Cottrell, G. W. NPClassifier: A Deep Neural Network-Based Structural Classification Tool for Natural Products. Journal of Natural Products. American Chemical Society November 26, 2021, pp 2795–2807. https://doi.org/10.1021/acs.jnatprod.1c00399.

Ludwig, M.; Nothias, L. F.; Dührkop, K.; Koester, I.; Fleischauer, M.; Hoffmann, M. A.; Petras, D.; Vargas, F.; Morsy, M.; Aluwihare, L.; Dorrestein, P. C.; Böcker, S. Database-Independent Molecular Formula Annotation Using Gibbs Sampling through ZODIAC. Nat Mach Intell 2020, 2 (10), 629–641. https://doi.org/10.1038/s42256-020-00234-6.

Nothias, L.-F.; Petras, D.; Schmid, R.; Dührkop, K.; Rainer, J.; Sarvepalli, A.; Protsyuk, I.; Ernst, M.; Tsugawa, H.; Fleischauer, M.; Aicheler, F.; Aksenov, A. A.; Alka, O.; Allard, P.-M.; Barsch, A.; Cachet, X.; Caraballo-Rodriguez, A. M.; Da Silva, R. R.; Dang, T.; Garg, N.; Gauglitz, J. M.; Gurevich, A.; Isaac, G.; Jarmusch, A. K.; Kameník, Z.; Kang, K. Bin; Kessler, N.; Koester, I.; Korf, A.; Le Gouellec, A.; Ludwig, M.; Martin H., C.; McCall, L.-I.; McSayles, J.; Meyer, S. W.; Mohimani, H.; Morsy, M.; Moyne, O.; Neumann, S.; Neuweger, H.; Nguyen, N. H.; Nothias-Esposito, M.; Paolini, J.; Phelan, V. V.; Pluskal, T.; Quinn, R. A.; Rogers, S.; Shrestha, B.; Tripathi, A.; van der Hooft, J. J. J.; Vargas, F.; Weldon, K. C.; Witting, M.; Yang, H.; Zhang, Z.; Zubeil, F.; Kohlbacher, O.; Böcker, S.; Alexandrov, T.; Bandeira, N.; Wang, M.; Dorrestein, P. C. Feature-Based Molecular Networking in the GNPS Analysis Environment. Nat Methods 2020, 17 (9), 905–908. https://doi.org/10.1038/s41592-020-0933-6.

Schmid, R.; Heuckeroth, S.; Korf, A.; Smirnov, A.; Myers, O.; Dyrlund, T. S.; Bushuiev, R.; Murray, K. J.; Hoffmann, N.; Lu, M.; Sarvepalli, A.; Zhang, Z.; Fleischauer, M.; Dührkop, K.; Wesner, M.; Hoogstra, S. J.; Rudt, E.; Mokshyna, O.; Brungs, C.; Ponomarov, K.; Mutabdžija, L.; Damiani, T.; Pudney, C. J.; Earll, M.; Helmer, P. O.; Fallon, T. R.; Schulze, T.; Rivas-Ubach, A.; Bilbao, A.; Richter, H.; Nothias, L. F.; Wang, M.; Orešič, M.; Weng, J. K.; Böcker, S.; Jeibmann, A.; Hayen, H.; Karst, U.; Dorrestein, P. C.; Petras, D.; Du, X.; Pluskal, T. Integrative Analysis of Multimodal Mass Spectrometry Data in MZmine 3. Nature Biotechnology. Nature Research April 1, 2023, pp 447–449. https://doi.org/10.1038/s41587-023-01690-2.

Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N. S.; Wang, J. T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T. Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res 2003, 13 (11), 2498–2504. https://doi.org/10.1101/gr.1239303.

Swift, C. L.; Louie, K. B.; Bowen, B. P.; Olson, H. M.; Purvine, S. O.; Salamov, A.; Mondo, S. J.; Solomon, K. V.; Wright, A. T.; Northen, T. R.; Grigoriev, I. V.; Keller, N. P.; O’Malley, M. A. Anaerobic Gut Fungi Are an Untapped Reservoir of Natural Products. PNAS 2021, 118 (18), 1–10. https://doi.org/10.1073/pnas.2019855118.

Wang, M.; Carver, J. J.; Phelan, V. V.; Sanchez, L. M. ; Garg, N. ; Peng, Y. ; Nguyen, D. D. ; Watrous, J. ; Kapono, C. A. ; Luzzatto-Knaan, T. ; Porto, C. ; Bouslimani, A. ; Melnik, A. V. ; Meehan, M. J. ; Liu, W.-T. ; Crüsemann, M. ; Boudreau, P. D. ; Esquenazi, E. ; Sandoval-Calderon, M. ; Kersten, R. D. ; Pace, L. A. ; Quinn, R. A. ; Duncan, K. R. ; Hsu, C.-C. ; Floros, D. J. ; Gavilan, R. G. ; Kleigrewe, K. ; Northen, T. ; Dutton, R. J. ; Parrot, D. ; Carlson, E. E. ; Aigle, B. ; Michelsen, C. F. ; Jelsbak, L. ; Sohlenkamp, C. ; Pevzner, P. ; Edlund, A. ; McLean, J. ; Piel, J. ; Murphy, B. T. ; Gerwick, L. ; Liaw, C.-C. ; Yang, Y.-L. ; Humpf, H.-U. ; Maansson, M. ; Keyzers, R. A. ; Sims, A. C. ; Johnson, A. R. ; Sidebottom, A. M. ; Sedio, B. E. ; Klitgaard, A. ; Larson, C. B. ; Boya P, C. A. ; TorresMendoza, D. ; Gonzalez, D. J. ; Silva, D. B. ; Marques, L. M. ; Demarque, D. P. ; Pociute, E. ; O’Neill, E. C. ; Briand, E. ; Helfrich, E. J. N. ; Granatosky, E. A. ; Glukhov, E. ; Ryffel, F. ; Houson, H. ; Mohimani, H. ; Kharbush, J. J. ; Zeng, Y. ; Vorholt, J. A. ; Kurita, K. L. ; Charusanti, P. ; McPhail, K. L. ; Nielsen, K. F. ; Vuong, L. ; Elfeki, M. ; Traxler, M. F. ; Engene, N. ; Koyama, N. ; Vining, O. B. ; Baric, R. ; Silva, R. R. ; Mascuch, S. J. ; Tomasi, S. ; Jenkins, S. ; Macherla, V. ; Hoffman, T. ; Agarwal, V. ; Williams, P. G. ; Dai, J. ; Neupane, R. ; Gurr, J. ; Rodríguez, A. M. C. ; Lamsa, A. ; Zhang, C. ; Dorrestein, K. ; Duggan, B. M. ; Almaliti, J. ; Allard, P.-M. ; Phapale, P. ; Nothias, L.-F. ; Alexandrov, T. ; Litaudon, M. ; Wolfender, J.-L. ; Kyle, J. E. ; Metz, T. O. ; Peryea, T. ; Nguyen, D.-T. ; VanLeer, D. ; Shinn, P. ; Jadhav, A. ; Müller, R. ; Waters, K. M. ; Shi, W. ; Liu, X. ; Zhang, L. ; Knight, R. ; Jensen, P. R. ; Palsson, B. Ø. ; Pogliano, K. ; Linington, R. G. ; Gutierrez, M. ; Lopes, N. P. ; ́; Gerwick, W. H. ; Moore, B. S. ; Dorrestein, P. C. ; Bandeira, N. Sharing and Community Curation of Mass Spectrometry Data with GNPS. Nat Biotechnol 2017, 34 (8), 828–837. https://doi.org/10.1038/nbt.3597.Sharing.

Xia, J.; Wishart, D. S. Web-Based Inference of Biological Patterns, Functions and Pathways from Metabolomic Data Using MetaboAnalyst. Nat Protoc 2011, 6 (6), 743–760. https://doi.org/10.1038/nprot.2011.319.