Best method for validating DNA methylation array data - (Apr/28/2010 )

Hi,

I am hoping someone on the forum may be able to provide me with some help on this... We have used the agilent dna methylation array to look at global DNA methylation in two different groups of patient samples. From this we found a number of probes on the array corresponding to genes that are differentially methylated.

In your experience what is the best way to validate differential methylation from methylation arrays? I realise that bisulfite sequencing would be the best way to do this. But based in the number of samples that will need to be examined x the number of genes we would like to validate this would take way more time than I have.... And I am the only one in the lab working on this...

So instead I have been looking at designing MSP primers. The problem with this is that the probes from the array are only 45 bp in length, are 100 bp apart in the genome and they don't overlap. As the MSP products need to be approx 150 bp in length, this only allows me to design either the forward or the reverse primer with information on the differential methylation status. For the opposite primer I designed it in CG dinucleotides but I don't know if these are methylated or not.... So I am taking a chance on this working...

Has anyone any experience/suggestions on the best way to deal with this? Also, I know that directly sequencing PCR products after bisulfite treatment is difficult. Does anyone have this method working well??

Thanks for any help!!

I think everyone in the field has the same issue about validating DNA methylation results.

I take it this is from an enrichment and interrogation on tiling arrays?

One other method I would suggest is to use another microarray platform to verify what you are seeing, such as the Illumina Infinium array, you can see how many of it's probes overlap with your hits and if a good proportion of them do, then I would invest in this effort. You will get a good snapshot of what is going on (27,578 CpGs) in the genome and correlate that with your current array results.