How to caclulate ChIP-Seq coverage?

Question

[Tori]

I have a ChIP-seq sample along with input control. I want to know the percentage of genome being enriched and coverage. I used bamFingerprint from deepTools and CHANCE. They gave conflicting results. How can I do it manuallly in a reliable way?

Answer 1

[Devon Ryan]

Presumably you have a BED or similar file with the position of the significant peaks. You can sum the widths of the peaks and divide by the total genome size to get the fraction of the genome enriched. For the coverage, samtools depth is a convenient enough method. Just average the per-site depths. You can also supply the BED file of your peaks to get the average coverage inside the peaks, if that's helpful.

[Tori]

I only have SAM file.

[Devon Ryan]

Then you need to use something to first call enriched regions. This is not something you want to do/write yourself unless you're into method development.

[liu4gre]

If I already have narrowPeak file, which contains the enriched region, how can I convert it to RPM or RPKM or tag density? Thanks.