Methylated DNA is over-represented in whole-genome bisulfite sequencing data

The development of whole-genome bisulfite sequencing (WGBS) has led to a number of exciting discoveries about how genomes utilize DNA methylation and has led to a plethora of novel testable hypotheses. Methods for constructing sodium bisulfite-converted and amplified libraries have recently excelled to the point that the bottleneck for experiments that use WGBS has shifted to data analysis and interpretation. Here we present empirical evidence for an over-representation of methylated DNA from WGBS. This enrichment for methylated DNA is exacerbated by higher cycles of PCR and is influenced by the type of uracil-insensitive DNA polymerase used for amplifying the sequencing library. Future efforts to computationally correct for this enrichment bias will be essential to increasing the accuracy of determining methylation levels for individual cytosines. MethylC-Seq of Arabidopsis thaliana