Tag size in MACS and MACS2

Question

[mozart.marins12]

Please, I need some clarification about using MACS in the Galaxy plataform. I have BAM files from my chip-seq experiment and the peak calling was performed with tag size 50. I have runned again the BAM files in MACS in the galaxy platform but using a tag size 36. The late gave me results that fit better to what I expect (less peaks in the control vs more peaks in the treatment). Is it ok to do that? I am very very raw in bioinformatics so I apolgize if this is a very basic question. Do I have to look the average of the tags resulting from the chip experiment?

Thank you

[mozart.marins12]

Ans is it possible somehow to change the tag size in MACS2 in the galaxy plataform?

Answer 1

[Ian]

I once asked the is question on the MACS forum (https://mail.google.com/mail/u/0/#inbox/148644b5135e70fb).

The answer from the author was: "Tag size only affects how MACS (version 1) builds strand model to compute fragment size. And in MACS2, it’s not even effective while computing fragment size since only ‘cutting’ positions are informative. But in MACS2, the so-called maximum gap (an internal value) for merging nearby significant regions is set as read length since we regard this as the resolution of your data. In fact, it has very little impact on peak calling. So… briefly, you don’t need to worry about this parameter. Longer reads help a lot for the reads alignment, but not much for peak calling."

I perform read trimming beforehand (MACS2) and the read lengths can be different. If you are not using MACS2 I would recommend it.