I have a generic question about MeDIP-seq analysis using MEDIPS bioconductor package. How many BAM files do I need for the analysis?

The example data "MEDIPSData" contains following BAM filles:

• hESCs.MeDIP.Rep1.chr22.bam(7.6M)
• hESCs.MeDIP.Rep2.chr22.bam(14M)
• hESCs.MeDIP.Rep3.chr22.bam(9.1M)
• hESCs.Input.chr22.bam(4.9M)
• DE.MeDIP.Rep1.chr22.bam(12M)
• DE.MeDIP.Rep2.chr22.bam(14M)
• DE.MeDIP.Rep3.chr22.bam(13M)
• DE.Input.chr22.bam(11M)

Do I need to create (or have) a similar set of files? Generally, we have only 2 fastQ files - R1 and R2. Are replicates essentials? Can I skip the normalization step?

Also, what about coupling set. How should I go about it?