Counting CpG's per read, please check my logic.

[Heisman]

Simple question but I want to make sure I get this right. Basically I'm looking at pulldowns of regions enriched with methylated CpG's.

Let's say I have a read that has the sequence: ACGTCGTACGCGTGCAAGCGC

The "CG" sequence appears 5 times.
The "GC" sequence appears 4 times.
However, there are only 7 distinct occurrences of "CG" or "GC" as one "CG" overlaps "GCGC" and one "GC" overlaps "CGCG".

A read may overlap the forward or reverse strand but either way CG will appear as the reverse complement of CG is CG.

Thus, I should only consider sequences of C followed by G, and not G followed by C, correct? Hence, in the above read, it will cover 5 CpG sites?

Is this correct logic?

[C.R.]

I would say your are correct. If you focus on CpGs then you should stay with the CG which is the same in forward and reverse orientation.

[mastal]

Is this correct logic?

Yes, it is correct.

[dpryan]

I'll add to the choir in saying, yes, that's correct. A G followed by a C is a GpC, not a CpG. A motif such as GCAA (i.e., a GpC or a CHH motif) generally has a very low rate of methylation, so you wouldn't expect it to contribute appreciably to the pull-down.

I imagine the fact that we focus on CpG sites and then talk about GC content can create a bit of confusion :P

[Heisman]

Excellent, thanks everyone.