How to construct co-expression network with correlation matrix?

Question

[MatthewP]

Hello, I know how to calculate pearson correlation matrix, but don't know how to construct co-expression network from it. Also shall I filter correlation matrix according each P value? Thanks!

Answer 1

[Arindam Ghosh]

Refer to Step-by-Step Construction of Gene Co-expression Networks from High-Throughput Arabidopsis RNA Sequencing Data. The section 3.12 Calculating Correlation of Gene Expression for Every Pair of Genes explains the steps.

Answer 2

[Clément Frainay]

I don't know much about co-expression networks, but maybe this can helps you : https://medium.com/@JavierBurroni/visualising-correlations-using-graph-2169c6415427

You can build a network with the igraph library (available python or R), from an adjacency matrix obtained by simply applying a threshold on your correlation matrix.

However you obtain a correlation network, not sure if this is enough to call this a co-expression network

[MatthewP]

Thanks, Clément Frainay. I know igraph but I have the same confusion with you that is this the co-expression network? Many papers said use pearson correlation to construct such network, but no detail how to perform. As for me, I want to construct co-expression only for differential expression genes and find hub genes(However I don't know such analysis pipeline is suitable or not). I know we can construct networks and find hub genes use gene list and string, but that is PPI, not co-expression network.

[Clément Frainay]

Ok. If you want to compare two conditions, maybe you want to look at differential networks? Not sure if it helps, but here is an example with detailed computation procedure : https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-11-95

[MatthewP]

Ok, I think this post is helpful for me. Thanks again!

[Kevin Blighe]

Hey Matthew, you mentioned my tutorial. Please let me know if you have issues in running it. I believe I show how to determine hub genes in the tutorial.