ranting about methylation analysis - (Dec/10/2004 )
I'm hoping someone can give me encouragement regarding why I am even pursuing my project. I was assigned to study the methylation status of two particular bidirectional promoters with CpG islands, but the more research I do, the more futile this seems. A full 14% of genes in the genome are bidirectional promoters with CpG islands and there is NO evidence that methylation is an important regulatory factor. MSP is frought with artifacts and whenever I read posts about that on this forum I just want to tell the poster not to do MSP, even though it's the simplest method (trust me, I've tried all of them). All of the original papers study cell lines, a HUGE problem as they are absolutely not the same as real cells, and the papers that use DNA from tumors only find methylation in fewer than 10% of their cases. Why am I even doing this? Why are we all doing this? What do we hope to find and how is it going to help the human race?
I share your concern.
I have no much knowledge about bidirectional promoters and how methylation may control their transcription. I think we should re-look at DNA methylation from a farer distance to put it into the picture of chromatin structure. Increasing evidence shows that there are interplays among DNA methylation, histone modification and even small RNA. DNA methylation change may be just an event secondary to histone modification and chromatin remodeling.
Regarding MSP, artifact is a problem especially if primers are not properly designed and PCR condistions are not optimized. To get something meaningful and serious especially for mechanistic studies, use the bisulfite genomic sequencing and cloning method.
Thanks pcrman, that's the method I am currently pursuing after trying several others. I'm trying to use the protocol that involves embedding the DNA in an agarose bead and am having trouble with the simple step of inverting the tube after the bead is made in oil-- everything gets stuck in the end of the tube and also the oil and water layers don't behave the way they should! It's truly strange. Does anyone have expertise with this protocol who can give me some advice? I know the next step is to just bite the bullet and buy a kit.