2
$\begingroup$

I am new in ATAC analysis.

I have a problem about how to transfer ATAC peak to gene symbol.

Here is the ATAC-seq: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM3271041

The GSM3271041_ATAC_sciCAR_A549_peak.txt.gz file contains five columns : (id, peak, chr, start, end). What I am doing is to use bedtools to transfer locations in chromosome to get gene symbols.

Here is an example from GSM3271041_ATAC_sciCAR_A549_peak.txt.gz

id,peak,chr,start,end
9,1-68591-69091,1,68591,69091
10,1-91105-91605,1,91105,91605

I use bedtools to get the gene symbol of id=9 , that is OR4F5.

I want to use these ''symbol-transfered'' ATAC-seq to integrate with RNA-seq which has gene symbols alreadly.

Am I right to do so ?

$\endgroup$
0

1 Answer 1

3
$\begingroup$

ATAC-seq data is fundamentally different from RNA-seq data because they measure different things. RNA-seq data is typically mapped from raw sequencing reads to annotated gene transcripts. This has a nice 1:1 correspondence because RNA is transcribed from genes.

ATAC-seq data, on the other hand, is used to measure accessible regions of the genome. But accessible regions don't come from genes the same way that RNA comes from genes. Accessible regions are scattered throughout the genome and most lie in intergenic regions where there is no annotation for a gene. Because accessible regions don't have the same biological connection to genes that RNA transcripts do, you need to be more precise in how you want to pair up accessible regions and genes.

Are you looking for cis-regulatory elements (CREs) that regulate nearby genes? Those will often be found within accessible regions, so you'd need additional data that connects an individual regulatory element to that gene.

Are you looking to create a catalogue of accessible regions within a certain cell type? Then you might want to look at ENCODE's candidate CREs database to see how your ATAC-seq peaks compares. From there, you'll be able to find additional information about those peaks, such as whether other functional genomics data identifies these regions as active regulatory elements, which gene(s) these elements regulate, and more.

There are different options depending on what exactly you're trying to do, but there is no one way to transfer ATAC-seq peaks to gene IDs because there's no biological question there.

$\endgroup$
2
  • $\begingroup$ Thanks for your answer! I think I want to map the peak in scATAC-seq to nearby genes. I found this. It's called peak annotation. $\endgroup$
    – Kevis Lin
    May 7 at 17:10
  • $\begingroup$ Annotating each with with the nearest gene(s) is a more specific question, that's great. Elsewhere on 10X's site you can see a description of their annotation process. Essentially, if you get a list of genes in a BED file, you can use bedtools closest with your peak list to find the nearest genes. The will usually work as a first pass for most applications, but like I said above, it will highly depend on your specific biological question. $\endgroup$ May 9 at 19:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.