Tools to plot strand cross correlation for a ChIP-seq experiment

[liz_is]

Hi all,

I'm looking for a tool which will allow me to plot the strand cross correlation of reads in some ChIp-seq data.

The estimate.mean.fraglen() function in the chipseq R package uses a correlation method to estimate fragment size, but gives me an answer far off what I'd expect - ~570bp rather than 200bp. The other methods the function can use give me ~200bp (SISSR) and ~380bp (coverage).

I'd like to plot the strand cross correlation to see why this might be happening. I've found this tool: http://code.google.com/p/phantompeakqualtools/, which is based on SPP (http://compbio.med.harvard.edu/Supplements/ChIP-seq/) but I am having trouble installing all the necessary dependencies on the lab server. Does anyone know of any alternatives? I don't need the peak-calling functionality of SPP, just the correlation plots.

Thank you!

[mudshark]

in principle you just need strand-specific coverage vectors of your ChIP signals and use the 'ccf' function in R to cross-correlate the two.


some remarks:
- why do you expect 570 bp?
- I once compared the true average fragment size obtained by paired-end sequencing and the one calculated by cross-correlation. Result: way different!

[liz_is]

I don't expect 570bp, I expect 200bp! Sorry if that wasn't clear.

I'll have a go with the ccf function, thanks.