This is a test version of Biostars. For the public version, visit https://www.biostars.org.
PeakNarrow ATAC-Seq analysis

Hello biostars,

I am new to this type of analysis. I have IDR.NarrowPeak data and I would like to normalise it to FPKM using R. I read that the packages limma and DESeq are helpful but I still have problems creating the count matrix. Is there a step I am missing?

Any help/suggestion will be appreciated.

atac-seq narrowpeak
• 2,495 views • link
updated by ATpoint ATpoint • written by Spacebio Spacebio

The count matrix is created outside of DEseq, have a look at featureCounts (subread toolkit, outside of R) or the countOverlaps function from GenomicRanges, in case you want to do it in R (not recommended as it requires he loading of bam files into R which is a waste of memory, so better use featureCounts directly on the bam files in the disk).

• 0 views • link
written by ATpoint ATpoint

2 answers

You should try DiffBind, which uses DESeq and edgeR in the background and automates this process: https://bioconductor.org/packages/release/bioc/html/DiffBind.html

• 0 views • link
written by igor igor

The peak file is the template for the count matrix. You can e.g. use featureCounts to get the counts per peak and sample directly from the BAM files:

## 1. Transform narrowPeak to the featureCount SAF format:
awk 'OFS="\t" {print $1"_"$2+1"_"$3, $1, $2+1, $3, "+"}' in.narrowPeak > out.saf

## 2. Run fC:
featureCounts -a out.saf -F SAF -o countmatrix.txt *.bam

This countmatrix you can normalize with e.g. DESeq2 or edgeR or transformations such as vst and rlog. Stay away from naive methods such as R/FPKM. They are not suitable for inter-sample comparisons.

• 0 views • link
written by ATpoint ATpoint

Log in to answer this question.

More posts like this