Mutated CpG sites

[assafgrw]

Hello,

I was wondering whether mutated cytosines at CpG sites—such as those resulting in TG dinucleotides—are included in the calculation of beta values.
Specifically, if a sequencing read contains a sequence like NNNTGNNN, would it still be considered as covering the CpG site for the purpose of methylation analysis, assuming the read is properly aligned based on the flanking regions?

Assaf

[ArtRand]

Hello @assafgrw,

The short answer is no. If a read doesn't have a cytosine, the base modification model won't make a cytosine modification call on that read. You should see these reads in the $\text{N}_{\text{diff}}$ column. If you're seeing a consistent mismatch you may want to look closer into what's going on there. Happy to help debug if you have some browser shots.