Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses

Abstract

Multi-omics datasets can provide molecular insights beyond the sum of individual omics.Diverse tools have been recently developed to integrate such datasets, but there are limited strategies to systematically extract mechanistic hypotheses from them. Here, we present COSMOS (Causal Oriented Search of Multi-Omics Space), a method that integrates phosphoproteomics, transcriptomics, and metabolics datasets. COSMOS combines extensive prior knowledge of signaling, metabolic, and gene regulatory networks with computational methods to estimate activities of transcription factors and kinases as well as network-level causal reasoning. COSMOS provides mechanistic hypotheses for experimental observations across multi-omics datasets. We applied COSMOS to a dataset comprising transcriptomics, phosphoproteomics, and metabolomics data from healthy and cancerous tissue from nine renal cell carcinoma patients. We used COSMOS to generate novel hypotheses such as the impact of Androgen Receptor on nucleoside metabolism and the influence of the JAK-STAT pathway on propionyl coenzyme A production. We expect that our freely available method will be broadly useful to extract mechanistic insights from multi-omics studies.

A Detailed Catalogue of Multi-Omics Methodologies for Identification of Putative Biomarkers and Causal Molecular Networks in Translational Cancer Research

“…Consequently, the development of multi-omics data integration methodologies that incorporate such prior biological knowledge should be enhanced as well, in order to delineate more readable causal networks between the perturbed omics in cancer manifestation. COSMOS (Causal Oriented Search of Multi-Omics Space) for example integrates phosphoproteomics, transcriptomics, and metabolomics data sets with prior knowledge such as protein-protein interactions to create hypotheses about causal links between signaling kinase cascades, transcriptional factors and metabolites [ 198 ]. Finally, a researcher should strongly consider to follow specific protocols such as the FAIR guiding principles (findability, accessibility, interoperability, and reusability) [ 199 ] when publishing multi-omics cancer data, and to take into account important bioethics considerations when sharing cancer patient data [ 200 ].…”

“…Consequently, the development of multi-omics data integration methodologies that incorporate such prior biological knowledge should be enhanced as well, in order to delineate more readable causal networks between the perturbed omics in cancer manifestation. COSMOS (Causal Oriented Search of Multi-Omics Space) for example integrates phosphoproteomics, transcriptomics, and metabolomics data sets with prior knowledge such as protein-protein interactions to create hypotheses about causal links between signaling kinase cascades, transcriptional factors and metabolites [ 198 ].…”

ViralLink: An integrated workflow to investigate the effect of SARS-CoV-2 on intracellular signalling and regulatory pathways

“…In addition, integration of CARNIVAL could extend the workflow to permit network reconstruction without supplying upstream perturbations (in this case the viral-host protein interactions). Whilst not currently integrated due to data availability issues, the addition of phosphoproteomics data to the pathway propagation methods could improve the prediction of active pathways (Dugourd et al 2020) Alternatively, methods to predict protein activity based on transcriptional signatures, such as VIPER and PROGENy (Alvarez et al 2016;Schubert et al 2018) could be added to the workflow in addition to network diffusion methods to increase the confidence of pathway predictions. Finally, extension of the network to include additional regulatory molecule types (e.g.…”

ViralLink: An integrated workflow to investigate the effect of SARS-CoV-2 on intracellular signalling and regulatory pathways

“…In addition, integration of CARNIVAL could extend the workflow to permit network reconstruction without supplying upstream perturbations (in this case the viral-host protein interactions). Whilst not currently integrated due to data availability issues, the addition of phosphoproteomics data to the pathway propagation methods could improve the prediction of active pathways [ 62 ] Alternatively, methods to predict protein activity based on transcriptional signatures, such as VIPER and PROGENy [ 63 , 64 ] could be added to the workflow in addition to network diffusion methods to increase the confidence of pathway predictions. Finally, extension of the network to include additional regulatory molecule types (e.g.…”