Description

DNA methylation typically refers to the covalent bonding of a methyl group (-CH₃) to the 5’ carbon of a cytosine residue. These methylated cytosines, referred to as 5-mC, most often precede guanines in the DNA sequence and are referred to as CpG methylation (Rakyan et al. 2011). DNA methylation is a critical component of the regulatory mechanisms that assure proper genome function. It plays an important role in embryonic development, X chromosome inactivation, genomic imprinting, and defense against the activity and transposition of repeat elements. Methylation is catalyzed by DNA methyltransferase enzymes (DNMTs) (Jeltsch 2006; Denis et al. 2011). Dysregulation of DNA methylation via mutations in DMNTs, disease processes or from environmental exposures is associated with neurodevelopmental disorders, chronic disease and cancers (Greenberg and Bourc’his 2019).

Bisulfite pyrosequencing (PSQ) is one method commonly used for quantitative analysis of DNA methylation at single nucleotide level. First, DNA undergoes bisulfite modification, a chemical process that converts only unmethylated cytosines into uracils. Methylated cytosines are protected from bisulfite conversion, remain as cytosines, and thus are distinguishable from unmethylated cytosines following PCR amplification and sequencing (Bassil et al. 2013; Li et al. 2013). We previously demonstrated that this method is able to distinguish as little as 0.5% differences in DNA methylation (Nye et al. 2015). Its accuracy makes it an important tool for determining DNA methylation levels (Bassil et al. 2013; Kurdyukov and Bullock 2016). It can be used as a means to validate findings from other ‘omics technologies, including the Illumina BeadChips, methylation capture followed by sequencing, reduced representation bisulfite sequencing, and whole genome bisulfite sequencing (Roessler et al. 2012; Ivanov et al. 2013; Kulis et al. 2015; Murphy et al. 2018). Limitations of PSQ for this purpose include the need for a dedicated pyrosequencing instrument, relatively low throughput, and amplicon size limitations (generally less than 200 bp). Additionally, if there is a need to examine multiple regions of the genome, for example as might be done to assess methylation biomarker panels, designing and validating multiple assays can be costly and time consuming. Thus, new methods that offer large-scale quantitative methylation analysis at single nucleotide resolution are a pressing need.

High-throughput DNA methylation sequencing combines bisulfite conversion technology and next generation sequencing to allow for detection of 5-mC residues at single-base resolution at specific regions of interest. In the methodology used here, DNA is bisulfite converted and ligated to an adapter including a unique molecular identifier (UMI) to allow accurate methylation analysis. The regions of interest are then enriched using single primer extension (SPE) technology with a pool of primers specifically designed for the regions of interest (Xu et al. 2017). This technique allows for increased throughput detection of methylation simultaneously at multiple specific regions versus bisulfite pyrosequencing. The specific QIAGEN technology utilized in this study allows for the investigation of DNA methylation at multiple regions across the genome at once, many of which contain multiple CpG sites per region, allowing for detection at thousands of CpG sites concurrently. While the method utilized in this study was optimized for use on Illumina Sequencing platforms, libraries could potentially be generated for analyses on other high throughput sequencing platforms, including PacBio or Oxford Nanopore. However, the reliance on harsh bisulfite treatments that may degrade DNA in this technology may create limitations in the types of samples that can be sequenced and the necessary input amounts of DNA. The use of a bisulfite modified template with single primer extension could also introduce strand biases for specific methylation patterns during the enrichment process. Additionally, the comparability of this method to other targeted methods and its ability to detect changes in methylation have not been extensively examined.

Here we measured and compared DNA methylation at CpG sites in genes involved in neurodevelopment in umbilical cord blood from 24 individuals using bisulfite pyrosequencing and high throughput DNA methylation sequencing. Individual CpG sites in the genes MAGI2, NRXN3, GRIK4, and GABBR2 were selected for this analysis. Across these four CpG sites, there was strong and statistically significant correlation between the percent methylation obtained via bisulfite pyrosequencing (PSQ) and obtained via QIAseq Targeted Methyl Panel technologies (QMS) (Figure 1A). These four CpG sites exhibit quite variable inter-individual methylation levels, showing that QMS is capable of accurately detecting levels of methylation across a wide range of potential methylation values.

At the CpG sites analyzed, there were differences between the overall levels of methylation detected by the two technologies. The range of the differences in methylation between PSQ and QMS were 0.05%-10.8% for MAGI2, 2.7%-14.7% for NRXN3, 0.1%-14.6% for GRIK4 and 0.2%-13.3% for GABBR2. Across all four CpG sites, the average difference between the two technologies in detected methylation for each sample was 5.6%. The levels of methylation significantly differed between methods with QMS reporting higher overall levels of methylation at the CpG sites within GABBR2, GRIK4, and NRXN3 and slightly lower overall levels of methylation at MAGI2. However, the overall spreads of the methylation distributions for the samples analyzed were remarkably similar (Figure 1B).

The results of this study demonstrate the potential of QMS to serve as an accurate and more rapid alternative to bisulfite pyrosequencing, particularly for analysis of multiple regions. Further testing is needed to corroborate these results and to better characterize the accuracy and limitations of detection of the QMS method. Rigorous testing of standards with known levels of DNA methylation is needed to determine the discriminatory potential across the range of possible methylation values for QMS, and further testing to determine correlations between QMS and other quantitative methylation techniques should be undertaken.