Home

Pathway Analysis Tools

Name	Manually-drawn?	Map Source	Data overlay?	API	Species filter?	Friendly?	Free?
Metacyc	Yes	Varied	N	N	Y	3	Y
BioCyc	Unknown	Varied	Y	Y	Y	1	N
SMPDB	Yes	Not declared	N	N	N, HU only	8	Y
WikiPathways	Yes	User-submitted	N	Y	Y	10	Y
KEGG	Yes	Drawn by staff	Y	Y	Y	4	Y
iPath	Yes (from KEGG)	KEGG	Y	Y	Y	10	Y
Pathway Projector	Yes (from KEGG)	KEGG	Y	N	N	9	Y

• MetaCyc - Metabolomics-centric website with the probably the best quality pathway database available - different tiers of reliability. Dealbreakers: can't answer the basic question: "which pathways do my couple hundred metabolites-of-interest reside in? (only allows up to 4 compounds), almost no pathway interactivity, pathways only downloadable as postscript or .pdf, neither of which are very editable after downloading.
• SMPDB - Multi-omics website, supported by Wishart Lab. Actually very close to the perfect metabolomics-centric pathway analysis tool, except that it's only for homo sapiens. Allows you to:
  • Given a list of metabolites-of-interest, search which pathways appear to be the most impacted (ranked by a rather opaque "score". Contrast with pathway enrichment in MetaboAnalyst, where explicitly mentions that it uses overrepresentation analysis.)
  • Click-through and explore your pathway of choice to (e.g. click on each of the nodes to see full name, chemical information, structure, etc...)
• KEGG - Manually drawn pathways, in .kgml format, following KEGG specific IDs, nomenclature, and categorization (e.g. KEGG modules or KEGG maps). Disadvantage: extremely difficult to get .kgml to talk to anything else, e.g. Cytoscape or Vanted. Is the database behind:
  • KEGG Mapper - Given a set of compounds of interest with their associated KEGG Ids, identifies them on a KEGG Map.
  • iPath - A friendlier version of KEGG Mapper, with an API.

Special Mentions

• Pathway enrichment analysis (via MetaboAnalyst.ca) - given an AUC table of the compounds of 2 experimental groups, returns a list of significantly impacted KEGG pathways.
• Vanted/Cytoscape - downloaded software to draw graphs in. Has access to KEGG. Note that KEGG maps were written more to be human- than machine-readable, though - this limits how reliable any data-overlaying operations may be (e.g. showing bar charts on nodes). For example, see hsa001900.

Metabolomics Data-Preprocessing Tools

A non-exhaustive list of open source metabolomics tools. If they're not here, they're either (a) deprecated (b) obscure enough that other major platforms don't use them, or (c) so difficult to find that they're probably not commonly used, which is in itself a good indicator of usefulness.

Click on the title of each tool for more details (if more details are required/available), or refer to the sidebar on the right.

• AMDIS - point-and-click software that searches NIST for metabolite profiling. Does not do quantitation.

• Biosigner - R package. Uses some machine learning methods to discover molecules which effectively discriminate between case and condition. Paradoxically, if you know enough about ML to use this package effectively, you probably don't need it (not when you can use the more powerful caret). https://bioconductor.org/packages/release/bioc/html/biosigner.html.

• CAMERA - R package. Does a bunch of peak-annotation. However, I've heard repeated warnings about its efficacy, and it has Build:error on its Github readme.

• EL-MAVEN - Free, open-source, GUI software, available for Windows or Mac. Does peak alignment, grouping, automated annotation (with reference to an external database or uploaded compounds list), isotope detection, and basic statistical analysis.

• GC-Autofit - Web app. Identifies and quantifies spectra from GC-MS.

• MetFrag - Annotates peak data by looking up your online database of choice (e.g. KEGG, HMDB, etc). Available as a web app (GUI), command line tool, or R package.

• MetExplore - Web app for exploring networks (has APIs!). Looks like an online version of Cytoscape. This got used to write a pathway enrichment tool for human metabolite network for W4M, within PheNoMeNal. https://metexplore.toulouse.inra.fr/index.html/index.php

• MetaboAnalyst - Available as a web app or R package. Performs analysis on data that's already for analysis, with some data processing tools as well. A really well-established tool (i.e. it does what it says it does extremely well).

• MSnbase - R package for metabolomics and proteomics data processing. Strong overlap with xcms, so much so that if xcms is included, then there's no need for MSnbase.

• Mzmine 2 - A tool kit for mass-spectrometry data processing (e.g. peak normalization/identification/alignment) and analysis (e.g. plotting, PCA, clustering), with the main focus on LC-MS data.

• MS Dial v3.7 - "MS-DIAL is a universal program for untargeted metabolomics and lipidomics supporting any type of chromatography/mass spectrometry methods (GC/MS, GC-MS/MS, LC/MS, and LC-MS/MS etc.)." Does peak detection, deconvolution, alignment, and identification, and probably some downstream stats as well.

• msconvert (ProteoWizard) - converts metabolomics file formats into other metabolomics file formats, e.g. .d to .mzML.

• xmcs-online -"Preprocesses GCMS or LCMS data for high-throughput, untargeted analyte profiling". A web-based version of the xcms R package; probably has fewer capabilities.