Difference in number of peaks called

Question

[anu014]

Hi Biostars,

I have used macs2 for calling peaks in chip-seq data. Firstly, I called peaks for antibiotic-treated sample & wild-type sample individually. After I ran macs2 with antibiotic-treated sample as treatment group (-t) & wild-type sample as control group (-c). Lastly, I ran the tool using antibiotic-treated sample as control group (-c) & wild-type sample as treatment group (-t). I got following number of peaks:

Antibiotic-treated S :  1007
WT S: 100
Antibiotic-treated vs WT : 46
WT vs Antibiotic-treated : 90

I am not getting the stats. If 100 were picked up individually & 90 peaks are called in WT vs Antibiotic treated run, then in Antibiotic-treated vs WT , the remaining 10 peaks should be checked against peaks present in Antibiotic-treated group. Why am I getting 46 peaks?

command parameters: # Command line: callpeak -f BAMPE -g 6748384 -B --mfold 1 60 --broad --broad-cutoff 0.1 --bw 300 --buffer-size 100000 --qvalue 0.05 --slocal 1000 --llocal 10000 --cutoff-analysis --fe-cutoff 1.0

Thank you in advance! :)

Answer 1

[colindaven]

Shouldn't you be calling each against the Input (same chip-seq process, but not applying the antibody) for these data, not against WT or control ? Have you looked at the plotfingerprints method from Deeptools - in Galaxy ?