Cell signaling network

Question

[jperezdelamar]

Hello Biostars,

this might sound very stupid but I have problems understanding the way to compute a cell signaling network. How TFs are connected to pathways? How to start the cell signaling from a GRN?

I've been reading a lot but I don't know why nothing is coming to my mind/imagination.

Thanks in advance!

Answer 1

[Kevin Blighe]

Dear Sir,

Let us first distinguish curated 'pathways' and then 'networks'.

Pathways

A 'curated pathway', like the ones in KEGG, will have experimental validation work behind them, i.e., evidence that is used to create the pathway. In such a pathway for a TF, for example, ESR1 (oestrogen receptor or ER), the precise sequence of regulatory events will typically be known and validated experimentally. That is, it is known which genes are influenced by the TF, and which further genes are then influenced by these other genes that are initially influenced by the TF. Upstream and downstream regulators are typically also known, such as estradiol (E2) in relation to the oestrogen signalling pathway.

In this screenshot, note the locations of the upstream regulatory, E2, and also the locations of the TF itself, ER:

---

Networks

A 'network' may not have any experimental validation and can be formed by just creating a correlation matrix from a set of genes.

A bit of revision first before we explain a network: a TF is a protein that will bind to regions of DNA and then initiate transcription of target loci. Potent transcription factors, i.e., ones that bind to many regions in the genome and initiate transcription of many genes, include ESR1 and MYC. Thus, the heightened activity of the TF will positively correlate with the transcription of all target genes. If I created an experiment where I dosed oestrogen-responsive cells with estradiol (E2) (in vitro) and then conducted RNA-seq on these cells, I could then create a network of ESR1 and all genes that statistically significantly correlate with ESR1. In this in silico network, ESR1 would be at the center, and all other target genes would be connected to this in a radial fashion.

I was looking for a good network plot to prove what I'm saying, and have now found one courtesy of Miao et al., 2017

The TFs are the triangles, with target genes, positioned radially around each respective TF, are circles:
free photo hosting