So after executing the peak calling program like MACS2, I get the bed file with peak intervals of different length. Now my question is: what are some of the common steps to select the top sequences and set them to equal length before using MEME-chip? Are there any statistical justifications behind it?
I am aware that I can select the top 600 sequence based on the peak height (pile-up), but I also notice that some of the summits are not necessarily "centered" in the peaks. Some summits are even near the left/right end of the peak interval. So if I simply create a bed file of 500bp regions centered at the summit at each peak, I know that I may include regions that are not in the peaks' interval.
I am new to this but I can't seem to find answers to this question on the internet. I appreciate for your time.