chipseq data normalization

Question

[kanwarjag]

I have 4 chipseq data (Bam files) files from Input, control treated and knockout of a gene. Now while preparing these samples different amount of DNA was processed/ pooled, because of inherent experimental problems. I want to normalize all four samples to fixed value and then analyze these samples. How best/ tool/ software I can do this normalization? Is it possible to get bam/ sam files of such normalized data as I would like to use NGS plot to generate figures for binding pattern.

Thanks

Answer 1

[A. Domingues]

I am using deepTools for my ChIP (and RNA-seq) normalizations and plotting:

https://github.com/fidelram/deepTools/wiki/Normalizations

For instance you can normalize to sequencing depth and input. The output, however is bigwig, but since this format is integrate in the other deeptools, it is ok for most purposes.

Answer 2

[Ying W]

Look into using DBChIP which can normalize based on non-peak regions or DiffBind which can normalize based on effective or full library size. The DBChIP paper is a good start if you are unsure what these normalization methods are.

Answer 3

[Ming Tommy Tang]

you can normalize to the library size.

[kanwarjag]

Could you please explain a little more on how we normalize to library size. Is there a package/ tool?

[A. Domingues]

deeptools will do that for you: https://github.com/fidelram/deepTools/wiki/Normalizations#bamCoverage