Table 1 Experimental approaches for profiling genome-wide DNA methylation
Experimental approach | Strength | Weakness | Resolution | Quantitative nature | Cost | Examples | References |
|---|---|---|---|---|---|---|---|
CHARM | -Cost-effective -Interrogate CpG sites genome-wide irrespective of proximity to genes or CpG islands | -Moderate resolution -Limited to regions in proximity to enzymes’ recognition sites | – | Abundance | Low | CGI shores show alteration DNA methylation in colon cancer [36] | [34] |
MBDCap-Seq | -Cost-effective -Allow the detection of DMRs within highly CpG-dense regions and regions with lower CpG density -MBD proteins can discriminate 5mC from 5hmC -No mutation introduced -More sensitive than MeDIP in regions with higher CpG density | -Relatively low resolution -Biased toward hypermethylated regions | ~150Â bp | Abundance | Moderate | Confirmed previous known differentially methylated sites and discovered new differentially methylated loci in 3 isogenic colon cancer cell lines [38] | [37] |
MeDIP | -Cost-effective -No mutation introduced -Specific to 5mC/5hmC depending on the antibody specificity -More sensitive in regions with low CpG density than MBDCap-Seq | -Biased toward hypermethylated regions -Do not identify individual 5mC sites -Inability to predict absolute methylation level | ~100Â bp | Abundance | Moderate | MBDCap-seq shows higher genomic coverage than MeDIP-seq along with twice as many DMRs between colon cancer and adjacent normal cells [45] | |
Illumina’s Infinium Methylation assay | -Cost-effective -Do not require a large amount of input DNA | -Human sample only -Coverage is highly dependent on the array design -Substantial DNA degradation after bisulfite treatment | Single base | Abundance | Low | DNA methylation as a signature to surrogate different cord blood cell types [49] | [47] |
WGBS | Evaluate methylation state of almost every CpG sites | -High cost -Substantial DNA degradation after bisulfite treatment -Cannot discriminate between 5mC and 5hmC | Single base | Digital | High | Bulk methylation level of CpG/CHG/CHH of wild-type Arabidopsis and methyltransferase-deficient mutants [18] Genome-wide methylation pattern and site-specific methylation [18] Global demethylation in the endosperm compared to the embryo [55] | [52] |
RRBS | -High CGI coverage -High sensitivity -Cost-effective comparing to WGBS | -May exhibit a lack of coverage at intergenic and distal regulatory elements -Substantial DNA degradation after bisulfite treatment -Limited to regions in proximity to enzymes’ recognition sites -Cannot discriminate between 5mC and 5hmC | Single base | Digital | Moderate | The EWAS study integrating DNA methylation, gene expression, proteomics, metabolomics and clinical traits in 90 mouse inbred strains [62] | |
scWGBS | Able to study methylome intra-population distribution | -Low sequencing efficiency (~20 million reads typically required per cell) -Cannot discriminate between 5mC and 5hmC | Single base | Digital | High | Determining epigenomic cell-state dynamics in mouse pluripotent and differentiating cells [74] | |
scRRBS | -Highly sensitive -Can detect target CpG sites at high coverage with relatively low number of sequence reads | -Substantial DNA degradation after bisulfite treatment -Cannot discriminate between 5mC and 5hmC -Provide relatively poor coverage for imprinting loci | Single base | Digital | High | Profiling epigenomic dynamics of 1 million CpG sites during early embryonic development in ESCs [70] | [70] |
TAB-seq | Can distinguish 5hmC from 5mC | -Substantial DNA degradation after bisulfite treatment -Tet enzyme with low efficiency might leave methylated residues unconverted -High sequencing depth is required to detect 5hmC with low abundance | Single base | Digital | High | Profiling 5hmC distribution in 108 days human PGCs to reveal DNA demethylation [83] | [81] |