Entering edit mode
3 months ago
QX
▴
60
Hi all,
I am doing ATAC-seq analysis, where I want to make a matrix of all consensus peaks from all replicates in all conditions. Can I just call all of them in one call using macs2
?
macs2 callpeak --verbose 3 -t *.bed \
-f BEDPE \
-g ${macs_genomesize} -B \
-q ${macs_q} \
--extsize 200 --nomodel --shift -100 --nolambda \
--keep-dup all --outdir ${peak_all_BEDPE_dir} -n "peak_all" --call-summits
or do I need to perform intersect
for replicates and then merge
for all conditions?
what is a more properly to deal with multiple replicates and conditions
I would call peaks for individual replicates and conditions rather than pool them together. This will also help you check replicate agreement.
after calling individually, how can you call the consensus peaks for all conditions? can you share the command for that?
Fully agree to rfran010's reply. And, compared to intersect peaks for all replicates, I think you may set consensus peaks for each condition to intervals which are more than 500bp and identified as peaks by at least 50% replicates.
do you know how to set this 500 and 50%? is that a
macs2
option?They're two independent steps. You need to call peak for each replicates and then use
bedtools genomecov
to compute coverage of each interval.https://bedtools.readthedocs.io/en/latest/content/tools/genomecov.html