Statistical test or method to test relationship between nearby genes and CpG sites

Question

[user911111111]

Dear All ,

I have a list of CpG sites and the genes lying close to the CpG sites, the respective beta values and gene expression values as well. I had use GRanges to find the nearby pairs of CpGs and Genes. Next, I wanted to test association or relation between the methylation of a CpG sites and the expression of the gene it lies close to. I would've performed pearson correlation, but I was wondering if there are any other methods that someone of you can suggest?

I look forward to your response and thank you in advance,

Best, S

Answer 1

[i.sudbery]

I've used correlation for similar things in the past, but I'd be skeptical of the p-values not being inflated. I'd consider using some sort of randomisation to generate the p-values.

I've used https://www.bioconductor.org/packages/release/bioc/html/InTAD.html to do this sort of thing.

Answer 2

[yura.grabovska]

You should consider using something like DMRcate if you have methylation array data in the form of beta-values. It groups probes into regions of methylation and does statistical testing for differential methylation.