Schematic representation of CGI-seq. TEST: single cell being studied. Methylation control (MC): cells that are used to define the detectable CGIs or any other accessible methylated regions (AMRs), sharing the same genome as the TEST. For CGI-seq analysis, the genomic DNA is classified into three classes of sequences: Me-CGIs (representing methylated CGIs, and more generally, Me-AMRs) indicated as a red line, Um-CGIs (unmethylated CGIs, and generally Um-AMRs) indicated as a blue line and undetectable regions indicated as a black line. Green lines are amplified sequences. Solid dot: Me-CpG sites. Hollow dot: Um-CpG sites. The intact gDNA is released from cell (step 1) and digested with RE1 (TEST) or no digestion (MC) (step 2), followed by MDA amplification (step 3). The amplification product is then digested by RE2 (step 4), in which the short fragments are converted t‘o a library and subjected to massive sequencing (step 5). The reads are aligned to the genome (step 6) and the methylation status of the AMRs (particularly CGIs) is determined (‘Materials and Methods’ section). The AMRs detected in both the TEST and the MC are called Me-CGIs (or Me-AMRs). The AMRs uniquely detected in the MC but not in the TEST are called Um-CGIs (or Um-AMRs) of the TEST. RE1 (four MREs in combination: Cfr42I (SacII), PdiI (NaeI), Eco52I (EagI) and PteI(BssHII), FastDigest enzyme from Thermal Scientific Inc.) distinguishes Me-CGIs (Me-AMRs) from Um-CGIs (Um-AMRs). MDA selectively depletes Um-CGIs (Um-AMRs) of the TEST and amplifies Me-CGIs (Me-AMRs). RE2 (either BstUI alone, or a combination of 3 separate REs (ABH: AciI, BstUI and Hinp1I from NEB)) generates fragments specifically enriching CGI sequences (and other CG-rich regions) from non-CGI sequences.