OmnipathR

An R client for the OmniPath web service and many other resources.

Package contents

  • Client for the OmniPath web service
  • Functions for post-processing OmniPath data
  • Access to other databases (18+ resources, see below)
  • Integration to NicheNet, a method to infer ligand activities from transcriptomics data

OmniPath web service client

OmnipathR retrieves the data from the OmniPath web service at

https://omnipathdb.org/

The web service implements a very simple REST style API. This package make requests by the HTTP protocol to retreive the data. Hence, fast Internet access is required for a proper use of OmnipathR.

What is OmniPath?

OmniPath is a database of:

  • Protein-protein, TF target and miRNA-mRNA interactions
  • Enzyme-PTM relationships
  • Protein complexes
  • Annotations of protein function, structure, localization, expression
  • Intercellular communication roles of proteins

To learn more about OmniPath, you can visit its website, or read our recent publication or our first paper from 2016, especially its supplementary material.

Access to further resources

The package provides access to a number of other databases: BioPlex, ConsensusPathDB, EVEX, Gene Ontology, Guide to Pharmacology (IUPHAR/BPS), Harmonizome, HTRIdb, Human Phenotype Ontology, InWeb InBioMap, KEGG Pathway, Pathway Commons, PrePPI, Ramilowski et al. 2015, RegNetwork, ReMap, TF census, TRRUST and Vinayagam et al. 2011.

Documentation

The latest version of the reference manual is available from https://static.omnipathdb.org/omnipathr_manual.pdf. Tutorials can be found at https://workflows.omnipathdb.org/. Sroll down for quick start examples.

OmniPath query types

We provide here a brief summary about the data available through OmnipathR. OmnipathR provides access to 5 types of queries:

  1. Interactions: protein-protein interactions from different datasets.
  2. Enzyme-substrate: enzyme-PTM (post-translational modification) relationships.
  3. Complexes: comprehensive database of more than 22000 protein complexes.
  4. Annotations: large variety of data about proteins and complexes features.
  5. Intercell: information on the roles in inter-cellular signaling.

For a more detailed information, we recommend you to visit the following sites:

https://omnipathdb.org/

https://omnipathdb.org/info

https://github.com/saezlab/pypath/blob/master/webservice.rst

https://saezlab.github.io/OmnipathR/articles/omnipath_intro.html

Installation

First of all, you need a current version of R (https://r-project.org).

From Bioconductor

OmnipathR is a freely available package deposited on Bioconductor and Github: (https://bioconductor.org/, https://github.com/saezlab/OmnipathR).

You can install it by running the following commands on a R console:

if (!requireNamespace('BiocManager', quietly = TRUE))
    install.packages('BiocManager')

## Last release in Bioconductor
BiocManager::install('OmnipathR', version = '3.12')
## Development version with the lastest updates
BiocManager::install('OmnipathR', version = 'devel')

From github

We add new features to OmnipathR way more often than the Bioconductor release frequency. To make use of the recent developments, you can use devtools to install the package directly from github:

require(devtools)
install_github('saezlab/OmnipathR')

Getting started and some usage examples

To get started, we strongly recommend to read our main vignette in order to deal with the different types of queries and handle the data they return:

https://saezlab.github.io/OmnipathR/articles/omnipath_intro.html

You can also check the manual:

https://saezlab.github.io/OmnipathR/reference/index.html

In addition, we provide here some examples for a quick start:

library(OmnipathR)

Download human protein-protein interactions from the specified resources:

interactions <- import_omnipath_interactions(
    resources = c('SignaLink3', 'PhosphoSite', 'SIGNOR')
)

Download human enzyme-PTM relationships from the specified resources:

enzsub <- import_omnipath_enzsub(resources = c('PhosphoSite', 'SIGNOR'))

Convert both data frames into networks (igraph objects)

ptms_g = ptms_graph(ptms = enzsub)
OPI_g = interaction_graph(interactions = interactions)

Print some interactions in a nice format:

print_interactions(head(interactions))

          source interaction         target n_resources n_references
4    SRC (P12931)  ==( + )==> TRPV1 (Q8NER1)           9            6
2  PRKG1 (Q13976)  ==( - )==> TRPC6 (Q9Y210)           7            5
1  PRKG1 (Q13976)  ==( - )==> TRPC3 (Q13507)           9            2
5    LYN (P07948)  ==( + )==> TRPV4 (Q9HBA0)           9            2
6  PTPN1 (P18031)  ==( - )==> TRPV6 (Q9H1D0)           3            2
3 PRKACA (P17612)  ==( + )==> TRPV1 (Q8NER1)           6            1

Find interactions between a specific kinase and a specific substrate:

print_interactions(dplyr::filter(enzsub,enzyme_genesymbol=='MAP2K1',
  substrate_genesymbol=='MAPK3'))

           enzyme interaction           substrate    modification n_resources
1 MAP2K1 (Q02750)       ====> MAPK3_Y204 (P27361) phosphorylation           8
2 MAP2K1 (Q02750)       ====> MAPK3_T202 (P27361) phosphorylation           8
3 MAP2K1 (Q02750)       ====> MAPK3_Y210 (P27361) phosphorylation           2
4 MAP2K1 (Q02750)       ====> MAPK3_T207 (P27361) phosphorylation           2

Find shortest paths on the directed network between proteins:

print_path_es(shortest_paths(OPI_g,from = 'TYRO3',to = 'STAT3',
    output = 'epath')$epath[[1]],OPI_g)

           source interaction          target n_resources n_references
1  TYRO3 (Q06418)  ==( ? )==>   AKT1 (P31749)           2            0
2   AKT1 (P31749)  ==( - )==> DAB2IP (Q5VWQ8)           3            1
3 DAB2IP (Q5VWQ8)  ==( - )==>  STAT3 (P40763)           1            1

Find all shortest paths between proteins:

print_path_vs(all_shortest_paths(OPI_g,from = 'DYRK2',to = 'MAPKAPK2')$res,OPI_g)
Pathway 1: DYRK2 -> TBK1 -> NFKB1 -> MAP3K8 -> MAPK3 -> MAPKAPK2
Pathway 2: DYRK2 -> TBK1 -> AKT3 -> MAP3K8 -> MAPK3 -> MAPKAPK2
Pathway 3: DYRK2 -> TBK1 -> AKT2 -> MAP3K8 -> MAPK3 -> MAPKAPK2
Pathway 4: DYRK2 -> TBK1 -> AKT1 -> MAP3K8 -> MAPK3 -> MAPKAPK2
Pathway 5: DYRK2 -> TBK1 -> AKT3 -> PEA15 -> MAPK3 -> MAPKAPK2
Pathway 6: DYRK2 -> TBK1 -> AKT2 -> PEA15 -> MAPK3 -> MAPKAPK2
.....

Alternatives

Python

A similar web service client is available for Python:

<https://github.com/saezlab/omnipath>

Cytoscape

The OmniPath Cytoscape app provides access to the interactions, enzyme-PTM relationships and some of the annotations:

<https://apps.cytoscape.org/apps/omnipath>

Customization

The pypath Python module is a tool for building the OmniPath databases in a fully customizable way. We recommend to use pypath if you want to:

  • Tailor the database building to your needs
  • Include resources not available in the public web service
  • Use the rich Python APIs available for the database objects
  • Make sure the data from the original sources is the most up-to-date
  • Use the methods in pypath.inputs to download data from resources
  • Use the various extra tools in pypath.utils, e.g. for identifier translation, homology translation, querying Gene Ontology, working with protein sequences, processing BioPAX, etc.

With pypath it’s also possible to run your own web service and serve your custom databases to the OmnipathR R client and the omnipath Python cient.

Feedbacks, bug reports, features

Feedbacks and bugreports are always very welcome!

Please use the Github issue page to report bugs or for questions:

https://github.com/saezlab/OmnipathR/issues

Many thanks for using OmnipathR!

License

Citation

Developers

  • Alberto Valdeolivas
    Author
  • Denes Turei
    Maintainer, author
  • Attila Gabor
    Author