Skip to content

COSMOS Prior-Knowledge Network

This vignette shows how to use omnipath-client to access the COSMOS prior-knowledge network (PKN) -- a multi-layer network designed for multi-omics causal reasoning with cosmosR.

The COSMOS PKN integrates metabolite-protein interactions, protein-protein signaling, and gene regulation from curated databases into a single directed signed network. It is served by the metabo.omnipathdb.org web service and built by the omnipath-metabo package.

Setup

pip install omnipath-client

import omnipath_client as oc

Network categories

The PKN is organized into six categories covering different layers of molecular interaction:

Category Description
transporters Metabolite transport across membranes (TCDB, SLC, GEMs, MRCLinksDB)
receptors Metabolite-receptor binding at the cell surface (STITCH, MRCLinksDB)
allosteric Allosteric regulation of enzymes by metabolites (BRENDA, STITCH)
enzyme_metabolite Enzyme-metabolite stoichiometric reactions from GEMs
ppi Protein-protein signaling from OmniPath
grn Gene regulatory network from CollecTRI / DoRothEA

Fetch the full PKN

# Full human PKN (all 6 categories)
df = oc.cosmos.get_pkn('human')

The result is a DataFrame with columns source, target, and sign (mode of regulation: 1 = stimulation, -1 = inhibition).

Filter by category

# Only transporters and receptors
df = oc.cosmos.get_pkn('human', categories=['transporters', 'receptors'])

# PPI layer only
df = oc.cosmos.get_pkn('human', categories='ppi')

Filter by resource

# Only STITCH receptor edges
df = oc.cosmos.get_pkn(
    'human',
    categories='receptors',
    resources='STITCH',
)

Multi-organism support

The PKN is available for human, mouse, and rat. The organism parameter accepts NCBI taxonomy IDs, common names, Latin names, or short codes.

# Mouse PKN
df_mouse = oc.cosmos.get_pkn('mouse')

# Rat PKN
df_rat = oc.cosmos.get_pkn('rat')

# Using NCBI taxonomy ID
df = oc.cosmos.get_pkn(10090)

AnnNet graph conversion

Convert the PKN to an AnnNet graph object for network analysis:

# Directly as an AnnNet graph
g = oc.cosmos.get_pkn('mouse', format='annnet')

# Or convert an existing DataFrame
df = oc.cosmos.get_pkn('human')
g = oc.cosmos.to_annnet(df)

Convenience functions

Explore what is available on the server:

# Available categories
oc.cosmos.categories()
# ['transporters', 'receptors', 'allosteric', 'enzyme_metabolite', 'ppi', 'grn']

# Supported organisms
oc.cosmos.organisms()
# [9606, 10090, 10116]

# Resources per category
oc.cosmos.resources()
# {'transporters': ['TCDB', 'SLC', 'GEM:Human-GEM', ...], ...}

Web service

The COSMOS PKN is served at metabo.omnipathdb.org. You can also query the REST API directly:

GET https://metabo.omnipathdb.org/cosmos/pkn?organism=9606&categories=all
GET https://metabo.omnipathdb.org/cosmos/pkn?organism=10090&categories=transporters,ppi
GET https://metabo.omnipathdb.org/cosmos/categories
GET https://metabo.omnipathdb.org/cosmos/organisms
GET https://metabo.omnipathdb.org/cosmos/resources