Skip to content

omnipath-metabo

Integrated prior-knowledge for metabolomics and multi-omics analysis.

omnipath-metabo builds and serves the COSMOS prior-knowledge network (PKN) -- a multi-layer directed signed network for causal reasoning across metabolomics, proteomics, and transcriptomics data with cosmosR.

What is omnipath-metabo?

omnipath-metabo integrates curated databases of metabolite-protein interactions, protein-protein signaling, and gene regulation into a unified network. It covers:

  • Metabolite transport -- membrane transporters from TCDB, SLC, genome-scale metabolic models (GEMs), and MRCLinksDB
  • Receptor binding -- metabolite-receptor interactions from STITCH and MRCLinksDB
  • Allosteric regulation -- enzyme modulation by metabolites from BRENDA and STITCH
  • Enzyme-metabolite reactions -- stoichiometric edges from GEMs
  • Protein-protein signaling -- kinase-substrate and other signaling from OmniPath
  • Gene regulation -- transcription factor targets from CollecTRI and DoRothEA

Key features

  • 7 resources -- TCDB, SLC, GEMs, MRCLinksDB, STITCH, BRENDA, OmniPath
  • 6 network layers -- transporters, receptors, allosteric, enzyme-metabolite, PPI, GRN
  • Multi-organism -- human, mouse, rat, and more via orthology translation
  • Pre-built PKNs -- ready-to-use networks served via the web service
  • Web service -- REST API at metabo.omnipathdb.org
  • AnnNet support -- convert to AnnNet graph objects
  • CLI -- cosmos-pkn command for scripted exports

Quick example

Build the transporter layer locally

from omnipath_metabo.datasets import cosmos

transporters = cosmos.build_transporters()

Fetch the full PKN via the client

import omnipath_client as oc

df = oc.cosmos.get_pkn('human')

Learn more

Services

Service URL What it provides
OmniPath Metabo metabo.omnipathdb.org COSMOS PKN, metabolite-protein interactions
OmniPath Utils utils.omnipathdb.org ID translation, taxonomy, orthology
OmniPath Database dev.omnipathdb.org Interactions, annotations, complexes, ontology