Can I directly use MACS2 -t -c for ATACseq differential analysis on WT and KO sample?

Question

[jxinyi118]

I have ATACseq data from one WT sample and one KO sample, and want to find different peaks between them. Can I just use MACS2 -t -c for this purpose?

For example,

MACS2 callpeak -t ATACseq_KOsample -c ATACseq_WT_sample -f BED -n "$prefix" -g $genome -p $pval_thresh --shift $shiftsize  --extsize $smooth_window --nomodel -B --SPMR

According to the tutorial of MACS2, it firstly calculates λ from ChIPseq IP, then tests input against Possion distribution based on λ. The p-value generated is used for peaks detecting.

Do you think the same logic can be used for the comparison between two ATACseq samples? in my case, a WT mice and a KO mice.

Thanks a lot!

To clarify, I don't have replicates for WT/KO so far, which is a problem, but I will have them.

Answer 1

[Friederike]

It is technically possible to do this comparison, i.e. I don't think MACS2 would complain. However, MACS2 will only focus on enrichments in the treatment sample over the control sample, so any losses of the KO sample would be difficult to come by (e.g. peaks where the signal is lost).

If you're getting replicates, then MACS2 is not the way to go anyway; you should rather look into the R packages DiffBind or csaw.

[jxinyi118]

Thanks! That's really helpful. :)