bamCoverage - how BPM normalization was done?

Question

[Pei]

Hello!

When using bamCoverage to generate coverage track for our ChIP-seq data, I choose BPM as the normalization method. With this setting, I found that the bamCoverage divided raw read count number per bin by one common denominator, for example, 10.15 for one library.

Could anybody teach me how this denominator was derived for each library? I could'nt figure out after read the tutorial from deeptools - I guess it may be related to library size?

Thank you!

Patrick

Answer 1

[Pei]

I think the norm factor was computed by dividing the number of mapped reads in the input bam by 500000. for example, my one lib had 5075424 mapped reads, so the norm factor is 5075424/500000 = 10.150. And hence raw read coverage in every bin was divided by 10.150