Coverage and normalization

[aliciagallego]

Hi Eric,

I want to generate metagenes of different regiones based on same BAM file. After creating the metagene object, I normalize the coverage by RPKM and do the plot (code below). However, I wonder if the normalization step is performed gene by gene and not in the overall number of genes falling within the same region set.

For example, if a set of regions, "A", is constituted by 10 genes and a set of regions, "B", contains 100 genes, and genes in "A" and "B" are similarly expressed, is metagene2 going to generate metagenes with same coverages between "A" and "B"? Or "B" metagenes will show higher mean coverages since they include a higher number of genes?

Thanks in advance.

My code:

mg <- metagene2$new(regions = regions_AB, bam_files = bam_files, force_seqlevels=TRUE, assay='rnaseq', bin_count=50)

mg$group_coverages(normalization="RPM")

mg$produce_metagene(title = "Metagene regions A and B")

[ericfournier2]

Hi,

metagene's RPM normalization is on a BAM-file basis, so the specific regions you pass in do not affect the normalization step.

Cheers,
-Eric