genome methylation analysis/database - analyze all promoters for CpG islands (Nov/30/2009 )

I am interested in looking for CpG islands in all promoters in the genome simultaneously. I tried writing my own PERL script, based on the algorithm described for Methprimer and using it to search 5kb promoters genome-wide, but it gives totally different results from Methprimer. I think it might be an issue of how they define observed/expected.

In any case, I am now wondering if there is a program out there already, or perhaps just a list of promoters with CpG islands that I could use.

MrMeth on Dec 1 2009, 02:17 AM said:

Hi,

you stumbled on a well known issue in the genome wide DNA methylation field; "CpG islands"...if there is such a thing really... were originally defined by the genome annotation teams by sliding a few hundred basepair frame over the genome and computing the GC density, where you start or how large the frame is determines what will be called a CpG island; ergo there are almost as many CpG islands defined as there are genome wide papers.... Read for instance Bock et al about CpG island predictors, a paper about how to remedy this problem.

If you want something that will stand the critique of peer review, just go to the UCSC genome table browser and cross section the CpG island track with the promoter track. There will be few reviewers who will take offense about you doing that and using that list of promoters to measure.


Best