To get correlation of different ChIP-seq samples

Question

[chaudharyc61]

Hello everyone

I wanted to get the correlation of different ChIP data samples. I have made 10kb windows and got z-scored log2 values of each window to plot the data.

I was wondering, to get the correlation, should i use z-scored log2 values or just log2 values or direct values which i got per window ??

Thank you Chandan kumar

Answer 1

[Shalu Jhanwar]

An average ChIP-seq peak size ranges between 200-500bp depending upon either TF or histone ChIP. Therefore, I'd suggest not using such huge windows (10kb) for comparison purposes.

I'd suggest performing peak calling first on the samples and then compute the correlation based on normalized counts within consensus peaks that can be obtained after merging the individual peaks from ChIP-seq samples.

[colindaven]

Good advice. Also, deeptools (in Galaxy) is excellent software for exploring these comparisons.