Is it logical to compare the expression FPKM with the percent methylation data at genes?

Question

[maria2019]

HI,

I have methylation data (C frequency from Meth call-bismark) and Bulk RNA-seq FPKM (from RSEM) on a cell line.

I am only interested in comparing RNA-seq and methylation relations at some specific genes on this cell line. ( I do not have any other control cell line - No Foldchange, FDR, etc).

Is it logical just to compare the expression FPKM with the mean percent methylation?

For example, I am interested in genes A, B, C

Reporst that I have:

               A    B   C

Methylation:  100   80  60

Expression:   300  40   250

-> Scores for each gene:

gene A: 300/100 = 3

gene B: 40/80 = ./5

gene C: 250/60= 4.16

Will this be a right comparison/scoring?

[ATpoint]

Depends on your question. Ff you want to check if gene methylation (gene body methylation) is correlated to its expression, then one might try to see if there is indeed a linear correlation between gene body methylation and expression. I did not follow the methylation field in the last years but I think there is indeed a connection, so I suggest to scan the literature for it. Let some papers inspire you about potential analysis strategies and ways to visualize this.

[maria2019]

Thanks for your comment. I have been reading different articles. There is for example an article (doi:10.1038/nature18606) that compares FPKM (log2) and Normalized RPKM (ATAC-seq data) - Figure 2a.

I am trying to find if anyone has done this based on methylation and how accurate it would be. Since for methylation I only have CpGs and I calculate the mean of CpG methylations at genes. I am not sure if that would be a fair comparison.