MACS2 peaks filtering

Question

[GK1610]

I have created a merged_BAM file by merging 100 individual bam files.

I am calling chip-seq peaks on merged_BAM file using MACS2

I want to obtain p-value threshold to filter out the MACS2 peaks

One of the ways I am doing is using IDR on 3 copies of merged_BAM that are created by randomly subsampling 33% of the original merged_BAM file.

merged_BAM_replicate 1( 33% of original merged_BAM) merged_BAM_replicate 2( 33% of original merged_BAM) merged_BAM_replicate 3( 33% of original merged_BAM)

Do you have other suggestions?

[ATpoint]

Why do you want to filter out peaks? MACS has internal q-value cutoffs, is there anything wrong with it? By subsampling, you reduce statistical power. Please explain what your goal is.