From research to clinical practice: DNA methylation signatures as diagnostic tools for Alzheimer's disease.

Based on their ability to analyze methylation differences genome-wide in a large number of samples at a reasonable cost and time, microarrays were rapidly adopted by Alzheimer's disease (AD) researchers to isolate CpG sites with methylation levels that were changed in the blood of AD patients. However, concerns gradually arose about the reproducibility of these methylation markers. I attribute the cause to the lack of confirmation of methylation changes detected by DNA methylation arrays. Only by simply plotting an individual's methylation distribution can some false methylation differences be avoided. Furthermore, we may need to set a realistic, minimum threshold for methylation changes for clinical purposes. The prospect of using DNA methylation as a diagnostic marker for AD may be compromised by the increasing number of irreproducible markers. Therefore, establishing minimum guidelines for the search and presentation of AD-associated methylation markers is necessary when publishing. Some empirically useful guidelines are outlined in this Perspective, and applications for diagnosing AD using reliable methylation markers are also presented. One of the most desired applications of methylation markers will be to identify individuals whose cognitive abilities decline before memory problems appear, a task that has not yet been accomplished with protein markers or imaging.