Entering edit mode
9.3 years ago
guiu
▴
60
Hello,
I am working on a Boolean network (GRN) I built from genes that are differentially expressed. Now I have to identify the key genes for the process I am studying. How do you go about that? Is there some "standard" pipeline for doing this?
- I know I can analyse the topology of the network, and rank the genes based on their metrics (centrality, degree etc). I can also find strongly connected components, modules etc. For this I know Cytoscape and its plug-ins/apps will do.
- Another approach would be to simulate the dynamics on the network, but I don't know what people usually do in this case. What is your preferred tool? And what information could I get from this simulation?
I have the feeling that people are mainly playing around with metrics and filters and keep the ones which give the results they want or something that is already known, instead of following a rigorous approach.
Thanks!
How exactly do you created your GRN? By co-expression in several condition? How do you know that a particular gene/protein regulates another gene/protein?
No i use interactions from a database (Metacore), so the interactions are manually curated from experiments, and also cathegorized by mechanism. I am using the differentially expressed transcription factors, getting their interactions from the database, and also contextualizing the resulting network to make it coherent with booleanized expression data.
What you should ask yourself first is: what biological question am I trying to address with this GRN ?
I want to find the drivers of this biological process, but just selecting the hubs of the network, or applying some selection on node metrics, just doesn't sound right to me. How can I do something that is biologically relevant, apart from "reverse engineering" my analysis from what I already know of the biological process?
You can run simulations and compare them to actual data. There are software to help you do this e.g. BioMdelAnalyzer and this paper for the description and this one for an example. Another, older, example is here. You may also be interested in this tutorial.