Correlation of DNA methylation and gene expression data sets

Question

[sinha.puja]

Hi everyone,

I have whole genome bisulfite sequencing (WGBS) data for methylation and RNA-seq data for gene expression. I needed to do correlation between these but could not find the correct way to proceed. I have two different files one for methylation and the other for gene expression. For methylation datasets I have methylation values (file containing columns with 30 subjects and rows representing chromosome number, position and gene ID) and for gene expression (file containing columns with 30 subjects and rows representing chromosome number, position and gene ID). Also the 30 subjects includes treatment and age effects. By searching I stumbled upon globaltest package in R but could not sure exactly whether to use or not. Does anyone help me in finding the better approach in R using some packages to do the analyses.

Thanks in advance.

Answer 1

[Kevin Blighe]

There are no right or wrong answers to this. If you have no pre-conceived hypotheses about the results, then you could apply WGCNA to each dataset separately, and then cross-compare results at the level of the derived WGCNA modules. Using WGCNA, you could also attempt to link different modules to treatment effects. I cannot guide specifically on this here - please follow the WGCNA tutorials for this.

You could also apply a generic network analysis ( see my own tutorial, here: Network plot from expression data in R using igraph ) and then merge the methylation and gene expression networks via igraph (it uses common nodes (genes) for the purposes of doing the merge).

For further generic guidance, see my pre-print with Dr. Leite: New insights in Tibial muscular dystrophy revealed by protein-protein interaction networks

Do search around, though... there are many posts on network analysis for what you are aiming to do.

Kevin