Table 2.

Characteristics of the experimental assays suitable for the analysis of DNA methylation in cell-free DNA samples, including sample requirement, DNA treatment protocols, summary of advantages and disadvantages of the assays, cost and references to relevant studies.

	Assay	Sample Requirement [Minimum]		DNA Treatment				Advantages	Disadvantages	Cost	Reference
				Bisulfite Conversion	Restriction Enzyme	DNA Precipitation	Enzymatic Plus Chemical Modification
Genome wide	Microarray	HM450	500 ng	■				pre-designed panel, easy to use, time efficient	DNA degradation; low coverage of intergenic regions	••	[21,77,78,147]
		HM850	250 ng	■				As above; includes enhancer regions; suitable for FFPE DNA;	As above	••
		MeKL-chip	10–20 ng			■		low DNA input	Cross hybridization, PCR amplification, MBD binding ability	••	[79]
	Whole genome sequencing	WGBS	10 ng	■				full methylome	DNA degradation; requires high sequencing depth; low input DNA may induce PCR bias	•••••	[51,53,54,56]
		PBAT	125 pg–10 ng	■				full methylome; PCR free; suitable for single cell analysis	DNA degradation; requires high sequencing depth; low fraction of aligned reads	•••••	[59,60,61,62]
		TAPS	1 ng				■	no DNA degradation; low input DNA; suitable for third generation sequencing; detect 5mC and 5hmC	hyper-active TET1 preparation	••••	[63,64]
		EM-seq	100 pg				■	no bisulfite DNA degradation; very low DNA input; high mapping; uniform GC coverage; detect 5mC and 5hmC	low complexity sequencing library	•••	[50,63]
	Representative genome wide methods	RRBS	10–100 ng	■	■			high CpG coverage	DNA degradation, low coverage of intergenic regions	•••	[66,67]
		scRRBS	one cell	■	■			very low DNA input		•••	[68,69]
		MCTA-seq	7.5 pg	■				very low DNA input	DNA degradation, low coverage of intergenic regions	•••	[70]
		cfMeDIP–seq	1–10 ng			■		no bisulfite DNA degradation, no mutation introduced, genome wide CpG and no CpG, very low DNA input	Detect only regions, low GpG density bias, CNA bias, depending on antibody performance	•••	[45,71,123,125]
		MBD-seq	5 ng			■		no bisulfite DNA degradation, outperform meDIP-seq in regions with higher CpG density	hypermethylated regions bias, CNA bias, depending on antibody performance	•••	[48,72]
Targeted	Sequencing	Target bisulfite seq	20-30 ng	■				high coverage on target loci	primer or probe design	••/•••	[58,80,81,82,83,84,85,122]
	PCR	MSPCR	Pg	■	■			low DNA in input; relative quantification of target loci	primer or probe design	•	[86,91,92,94,95,148]
		ddMSPCR	pg	■	■			low DNA input; absolute quantification of target loci	primer or probe design; quantification depends on DNA input	•	[97,98,99]

HM450: Infinium Human Methylation 450 Bead Chip (Illumina), HM850: Infinium Methylation EPIC BeadChip Kit (Illumina), MeKL-chip: methylated DNA, Kinase pre-treated ligation-mediated PCR amplification and hybridization to the CHARM array, WGBS: whole genome bisulfite sequencing, PBAT: post bisulfite adaptor tagging technique, TAPS: TET-assisted pyridine borane sequencing, EM-seq: Enzymatic Methyl-Seq, RRBS: Reduced representation of bisulfite sequencing, scRRBS: single cell reduced representation of bisulfite sequencing, MCTA-seq: Methylated CpG tandems amplification and sequencing, cfMeDIP–seq: circulating free methylated DNA immunoprecipitation sequencing, MBD-seq: methyl-CpG binding domain protein capture sequencing, Target bisulfite seq: target bisulfite sequencing, MSPCR: methylation specific PCR, ddMSPCR: droplet digital methylation specific PCR. The number of filled circles schematically represents the cost of the assays, from cheaper (•) to more expensive (•••••) assays.