An unbiased approach to measure aberrant DNA methylation alterations

The ability to accurately measure aberrant DNA methylation levels, which occur in multiple diseases including cancer, is integral to the understanding of DNA methylation biology. It is well-established that in cancer, the largest, and thus, most biologically important absolute gains of DNA methylation levels occur at CpG sites with low native levels while the largest losses occur at CpG sites with high native levels. Conventional wisdom assumes that the observed association between the degree of the alterations and the native levels are largely due to the limitations of change within the DNA methylation scale, i.e. the maximum possible gain/loss of DNA methylation depends on the native level. Here, we present evidence that this association is largely caused by alterations occurring as a global rate of change relative to the native level. We also present evidence that current approaches that measure alterations as an absolute level change have an implicit bias for the native level, which has likely skewed our understanding of DNA methylation biology. We found that DNA methylation alterations can be accurately compared by calculating the rate of change relative to the native level. Most importantly, this approach enables the identification of more biologically significant DNA methylation alterations.