Normalization methods for spiked Chip Seq data

Question

[MolGeek]

Hello Everyone.

I am currently analyzing spiked chip seq data and i am searching for an appropriate normalization method to get scaling factors, in order to create bigwigs.

I have the alignment percentages in both human genome and drosophila.

Any suggestion?

Thanks in Advance!

Answer 1

[ATpoint]

As in ATAC-seq sample normalization but the raw.counts matrix would simply be the matrix of raw counts with only the spike-ins.

[MolGeek]

Thats an interesting method. But can i avoid using deseq or edgeR matrices? Would be time consuming. Is there any method that doesnt include peak calling and downstream analysis, rather than bams or bigwigs?

[rfran010]

I think doing it this way would be more appropriate like you requested.

It shouldn't be too time consuming. Peak calling is relatively quick with macs2, and you can keep it simple by following up with deeptool's multiBamSummary BED-file with --scalingFactors option.

[MolGeek]

I will try this method. A few questions i have is:

Should the matrix be the macs output using the dm6 or the hg38 bams? Should i use the inputs in macs?
My chips are paired end. Is bamCoverage -b input.bam -o ouput.bw --scaleFactor xx -p 20 correct for paire end data?