What's "important" depends on the biological quesiton.
If your experiment is properly designed with replicates, you can do a differential analysis to find the peaks that are significantly differentially bound between conditions (eg different antibodies). The
DiffBind package in R/Bioconductor does this and has a step-by-step tutorial. You can also do simple overlaps (venn diagrams) to find peaks in common or peaks unique to a condition, but you can't get confidence statistics for this.