The state of the economy, global climate change, DNA methylation…These are all complicated issues. Well, we found one well-written review article by Thomas Mikeska, Ida Candiloro, and Alexander Dobrovic at the Peter MacCallum Cancer Centre and the University of Melbourne in Australia that tries to put one of those topics, DNA methylation in perspective, at least. Hopes have been pinned on using DNA methylation as a biomarker for various cancers, but as the authors caution in their article, DNA methylation isn’t quite so simple.
The most basic scenario occurs when all CpG sites on a set of alleles (known as epialleles) are unmethylated under normal conditions but become methylated in a disease. But DNA methylation isn’t always all or nothing. For instance, within a given set of epialleles, a few CpG sites on a DNA strand can be methylated, all sites on a different strand can be completely methylated, and another strand can sit untouched. Techniques to quantitate methylation, such as MethyLight, bisulfite pyrosequencing, or COBRA, are great at distinguishing fully methylated sites from unmethylated ones but lump partially methylated together with the fully methylated sites.
The authors propose that the direct visualization of individual clones is the best way to detect and quantify heterogeneously methylated epialleles. The most suitable approach is a limiting dilution digital method where single templates are amplified without PCR bias. But such an approach is expensive because multiple PCRs have to be done. However, as technologies improve, the authors are optimistic that analysis of multiple reactions will be more cost-effective and commonplace. Heterogeneously methylated regions can then be used cancer biomarkers.
See if reading the review will help un-complicate things at Epigenomics, August 2010