Hi all, I would like to ask how people normalise their data of BrdU-IP sequencing? This is a plot of BrdU-IP-seq across a region in chromosome V in yeast. BrdU labels nascent DNA so here I am basically seeing newly replicated DNA as the cells progress from early S phase to late S phase. As you can see, replication in yeast starts in defined sites called "origins". The replication machinery starts replicating from origins, afterwards it progress in both directions. This is why from early to late S phase, you see a spreading of the signal.

I analysed my data by 1) trimming my reads with TrimGalore, 2) aligning to yeast genome with Bowtie2, 3) filtering reads (i.e. remove multi mapped reads and duplicates with Samtools) then compute the genome coverage using BamCoverage to generate the bigWig. I normalise with RPKM

My question is this: At timepoint=20 mins, I get well defined signal in specific origins. I would expect the signal at those sites to keep increasing in succeeding timepoints, but the graph shows decreasing signal there. I guess because bamcoverage computes the coverage at each bin by taking into account the total reads. Because late replicating cells have more replicated regions then they have more total reads. How can I change my analysis to visualise my data such that I see increasing signal? Can I just not normalise my data?