Question

How to select one gene at a CpG site?

0

Entering edit mode

5.4 years ago

hansong798 ▴ 20

Hello, I have some problem to process TCGA public data.

I downloaded TCGA breast cancer's methylation data with annotation information using "TCGAbiolinks" package of R but there are some CpG sites which have multiple gene symbols at each.

However, I need data which have 1 gene symbol at 1 CpG site for analysis.

Is there any criteria or method?

enter image description here

gene assembly • 1.5k views

ADD COMMENT • link updated 3.1 years ago by wuhan9281 • 0 • written 5.4 years ago by hansong798 ▴ 20

0

Entering edit mode

Please provide an example site and the genes you see.

ADD REPLY • link 5.4 years ago by GenoMax 141k

0

Entering edit mode

I added figure. It is a part of my annotation data.

ADD REPLY • link 5.4 years ago by hansong798 ▴ 20

0

Entering edit mode

Hi! I am also puzzled with this problem. How did you deal with this problem at last?

ADD REPLY • link 3.1 years ago by wuhan9281 • 0

score 0 · Answer 1 · 2018-12-14

I think bi-directional promoters exist for a non-trivial number of genes.

While I don't think the overall percentage (or the gene-overlapping percentage) is as high as you may expect from your screenshot, one possible solution would be to use the gene regions like TSS200 or TSS1500 and only consider them if they only have one gene annotation (and/or keep the paired name, and visualize methylation and/or expression for the region including both genes).