COSMOS (Causal Oriented Search of Multi-Omic Space) is a method that integrates phosphoproteomics, transcriptomics, and metabolomics data sets. COSMOS leverages extensive prior knowledge of signaling pathways, metabolic networks, and gene regulation with computational methods to estimate activities of transcription factors and kinases as well as network-level causal reasoning. This pipeline can provide mechanistic explanations for experimental observations across multiple omic data sets.
COSMOS uses CARNIVAL’s Integer Linear Programming (ILP) optimization strategy to find the smallest coherent subnetwork causally connecting as many deregulated TFs, kinases/phosphatases and metabolites as possible. The subnetwork is extracted from a novel integrated PKN (available here) spanning signaling, transcriptional regulation and metabolism. Transcription factors activities are inferred from gene expression with DoRothEA, a meta resource of TF/target links. Kinase activities are inferred from phosphoproteomic with a kinase/substrate network of Omnipath, a meta resource of protein-protein. CARNIVAL was adapted to find mechanistic hypotheses connecting the TF and kinase activities with metabolites from a signaling/metabolic prior knowledge network combining Omnipath, STITCHdb and Recon3D. The scripts to generate the current network can be found here: https://github.com/saezlab/meta_PKN_BIGG.
You can also use COSMOS if you don't have metabolomic data, to connect TF activities (from transcriptomic) with kinase activities (from phosphoproteomic) for exmaple !
R >= 4.1 is required
# install from bioconductor
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("cosmosR")
# We advise to instal from github to get the latest version of the tool.
if (!requireNamespace("devtools", quietly = TRUE))
install.packages("devtools")
devtools::install_github("saezlab/cosmosR")If you don't have R 4.1, you can also clone the github repository on your machine, create a new R project with R studio from the cosmosR folder, change the R version to your own R version in the DESCRIPTION file and then install it with devtools:install()
But 4.1 is advised in any case.
Here you can find an extensive tutorial showing how to use MOFA and COSMOS with the NCI60 dataset. This is an extensive tutorial, if you wish to get a quicker plug and play introduction to COSMOS, see below.
We made a repository that contains pre-processed inputs and an example script to use cosmos with the NCI60 RNA+metabolomic datasets. You can find the repository here.
We recorded a video guide for cosmosR tutorial in the context of a course organised by EBI-EMBL. You can access the recording at this link for a step by step introduction to cosmosR : https://embl-ebi.cloud.panopto.eu/Panopto/Pages/Viewer.aspx?id=318f7091-b6bf-44ee-939f-adb10121fc1b
The meta PKN used with the biorXiv version of COSMOS is available here.
An updated meta PKN is available with the package (using data(meta_network) in R)
If you use cosmosR for your research please cite the original publication:
Dugourd A, Kuppe C, Sciacovelli M, Gjerga E, Gabor A, Emdal KB, Vieira V, Bekker-Jensen DB, Kranz J, Bindels EMJ, Jesper V Olsen, Christian Frezza, Rafael Kramann, Julio Saez-Rodriguez et al (2021) Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses. Mol Syst Biol 17: e9730
The code is distributed under the GNU General Public License v3.0. The meta PKN is distributed under the Attribution-NonCommercial 4.0 International (CC-BY-NC 4.0) License.