Cross correlation profile for Input data set (without IP)

[GoogleCodeExporter]

Dear all,

I'm using phantompeakqualtools to compute cross-correlation for a set of 
different histone modification ChIP-seq data and related input control with 
biological replicates.

Since it is the first time I am working on ChIP-seq data, I don't have yet 
enough background to be confident in graphes I obtained.

I would like to know :

1) Is it expected to see a peak corresponding to fragment length for the input 
data set (i.e. without IP) ?

2) Is the difference between cross-correlation profiles related to biological 
replicates sufficient to throw a biological replicate ? Actually, for some of 
the histone marks I am analyzing, profile as well as NSC and RNC are really 
different between replicates (an example is provided attached), I was wondering 
if I should still consider replicates with low NSC/RNC.

Thanks a lot for helping.

Best,

Pierre-François

Original issue reported on code.google.com by pierre-f...@orange.fr on 29 Apr 2015 at 2:16

Attachments:

• D6K4Me1_Rep1_PHANTOM.pdf
• D6K4Me1_Rep2_PHANTOM.pdf