I am working on analyzing CUT&RUN H3K27me3 for which we included spike-in mononucleosomes. I was planning on aligning our sequencing data to a reference genome including the spike-in sequences.
For normalization to spike-in sequences, I was planning on calling peaks, counting the number of reads on significant peaks, and generating scaling factors using DeSeq2 (dds <- estimateSizeFactors(dds, controlGenes= <names or numeric index of control features> ).
Subsequently, I am planning on normalizing bam files through the "bamcoverage" command in deepTools. I have two questions:
1) overall, does this seem like a reasonable approach for spike-in normalization? 2) when calling bamcoverage, should I normalize to RPKM and --scaleFactor, or normalize to just --scaleFactor? On the deep tools page, it seems like both approaches are viable, however, I wanted to see what the general consensus is.