macs2 call peak

Question

[yiren]

I used macs2 to call peaks for chip-seq data.But I do not know why macs2 will slide two bandwidth along the genome for calling peaks?

[ATpoint]

Did you read the paper?

[yiren]

yes,I read this paper(Model-based Analysis of ChIP-Seq (MACS)),but I i can not find why.did any paper explain this?

[ATpoint]

From what I understand it is simply a window to detect candidate peaks. By default 600bp., so 2BW with BW=300bp which is more or less a typical fragment length in standard ChIP-seq. Double that size because the actual peak should be larger than the fragment size (as a peak is not a perfectly horizontal stack of reads) so 2BW sounds reasonable to me.

Answer 1

[ben.kunfang]

I think it is because sequence tech only read the end of the reads, so the actual position of the peak is not the peak of the reading reads. In this case, macs2 will shift a distance to get the real peak position. please check "calling peak" part in this link https://galaxyproject.org/tutorials/chip/