Site specific Bisulfite analysis Problem - (May/22/2008 )
This forum has been a life saver for points in setting up my current bisulfite assay. So props to all the sponsors that make this work.
I've amplified and sequenced my promoter. The sequence looks good (no methylation whatsoever with 100% conversion of Cs). However, at a single G, ~50% of the signal becomes T after the bisulfite reaction. We've now performed ~30 samples in duplicate and it just keeps coming up. The G is in the sequence AGG_CACCA. Is there a rational reason why a G would convert to a T? And why would it only happen on the single G of the 200bp sequenced?
It may be possible the sample is heterozygous G/T at that location. Did you clone your PCR products or go straight to direct sequencing? If you used direct sequencing then it is possible, if you cloned then I can't explain it with my experience.
We direct sequenced. We did not clone. I should clarify that the unconverted DNA is only a strong G at this site. I do not believe that this is a SNP on the DNA, but something that is occurring as a part of the bisulfite treatment. Why it comes up at 50% intensity in the bisulfite converted DNA, but not the nonconverted DNA is really bothering me.
I've amplified and sequenced my promoter. The sequence looks good (no methylation whatsoever with 100% conversion of Cs). However, at a single G, ~50% of the signal becomes T after the bisulfite reaction. We've now performed ~30 samples in duplicate and it just keeps coming up. The G is in the sequence AGG_CACCA. Is there a rational reason why a G would convert to a T? And why would it only happen on the single G of the 200bp sequenced?
bisulfite shouldn't cause a transversion from a G to a T. If it is not a SNP, then the other possibilities are
1. misincorporation by the taq pol
2. sequencing artifact
to figure this out, you would have to clone the fragments. If half gives G and the other half give T, this would point to the PCR step, where the taq misincorporates a T for a G or an A for a C on the other strand.
don't know, but it doens't seem like a huge problem to me.