PT  - JOURNAL ARTICLE
AU  - Piran, Mehran
AU  - Karbalaee, Reza
AU  - Piran, Mehrdad
AU  - Mirzaie, Mehdi
AU  - Ansari-Pour, Naser
AU  - Tang, Jing
AU  - Jafari, Mohieddin
TI  - Do signaling networks and whole-transcriptome gene expression profiles orchestrate the same symphony?
AID  - 10.1101/643866
DP  - 2019 Jan 01
TA  - bioRxiv
PG  - 643866
4099  - http://biorxiv.org/content/early/2019/05/21/643866.short
4100  - http://biorxiv.org/content/early/2019/05/21/643866.full
AB  - Studying relationships among gene-product expression profiles is a common approach in systems biology. Many studies have generalized this subject to different levels of the central dogma information flow and assumed correlation of transcript and protein expression levels. All these efforts have updated the signaling network models and expanded the current signaling databases, which include interactions among the gene-products extracted based on either the literature or direct and indirect experiments. In fact, due to unavailability or high-cost of the experiments, most of the studies do not look for the direct interactions (gene-protein or protein-protein) and some of them are contradictory. In addition, it is now a standard practice to undertake enrichment analysis on biological annotations especially in omics research to make claims about the potentially implicated biological pathways in disease. Specifically, upon identifying differentially expressed genes, molecular mechanistic insights are proposed based on statistically enriched biological processes for disease etiology and drug discovery. However, it remains to be demonstrated that expression data may be used as a reliable source to infer causal relationships among gene pairs. In this study, using four common and comprehensive databases i.e. GEO, GDSC, KEGG, and OmniPath, we extracted all relevant gene expression data and all relationships among directly linked gene pairs in order to evaluate the rate of coherency or sign consistency. We illustrated that the signaling network was not more consistent or coherent with the measured expression profile compare to random relationships. Finally, we provided the pieces of evidence and concluded that gene-product expression data, especially at the transcript level, are not reliable or at least insufficient to infer biological relationships among genes and in turn describe cellular behavior.