| DNA methylation is an epigenetic mark at the interface of genetic and environmental factors relevant to human disease. Quantitative assessments of gene specific and global DNA methylation levels are thus important tools in epidemiology research, particularly for understanding the effects of environmental exposures in complex diseases. Among the available methods of quantitative DNA methylation measurements, bisulfite sequencing is considered the gold standard, but whole genome bisulfite sequencing (WGBS) has previously been considered too costly for epidemiology studies with large sample sizes. Part of this study used the more cost effective pyrosequencing to validate differential methylation identified by WGBS in human placenta and found pyrosequencing technology useful in loci specific epigenetic biomarker analysis. However, pyrosequencing of repetitive sequences within bisulfite treated DNA has also been routinely used as a surrogate for global DNA methylation. A comparison of pyrosequencing to WGBS for accuracy and reproducibility of global methylation levels has not been previously reported. Therefore, this study also compared the global methylation levels measured from unselected WGBS sequences to pyrosequencing assays of several repeat sequences and repeat assay-matched WGBS data. Sources of variation in repetitive pyrosequencing assays were determined to include PCR amplification bias, PCR primer selection bias of targeted sequences, and inherent variability in methylation levels of repeat sequences. Low coverage, unselected WGBS showed the strongest correlation between replicates of all assays. By utilizing indexed barcode multiplexing, the cost of WGBS can be lowered significantly to improve the accuracy of global DNA methylation assessments for human studies. |
