ATAC data analysis

Question

[1769mkc]

I am dooing atac seq data analysis so far what i have done

• I aligned the data with bowtie2
• Removed duplicates
• made bed files out of those bed file
• Did peak calling with Homer using it findPeak command -Now i have list of peak files with respect to various condition . In Homer manual i read a function called findMotifsgenome function , i have some confusion here how do i proceed from here do i have to just give the peak file as it is or I have to process the peak files .

Any suggestion or help would be highly appreciated

Answer 1

[Kasthuri]

Typically we use diffbind to find differentially binding sites between the control (closed chromatin) and experiment (open chromatin). I don't use Homer. But one can use bedtools getfasta to get the fasta sequences for those open chromatin peak sites and use MEME to get the consensus motif if that is what one wants. There are other analysis one can do, like using deepTools etc.

[1769mkc]

okay so i have to convert the peak files into fast and then use MEME? thats what you are suggesting

[Kasthuri]

Yes, that is one reasonable approach...

[1769mkc]

i have all the peak files but what do i need from the peak file as input to to do MEME becasue it contains field like peak ID chromosome location , fold change etc etc .I dont think i need all those .Do i have to extract some columns from my peak files or i just have to use the whole file?

[Kasthuri]

You need to have a bed file - just use the chromosome, start and end position and get the fasta sequences through bedtools and use MEME.