Figure 4

Identification of imprinted DMRs at the proximal promoter regions of FAM50B and MCCC1. (A) and (F) Schematics showing the positions of methylation assays (Biseq: bisulphite cloning and sequencing assay; cg code: probe number of Illumina assay; and Pyro: bisulphite pyrosequencing assay) and SNP locations relative to the genes. Arrow directions represent the transcriptional directions for the genes. Genomic coordinates were retrieved from the UCSC Genome Brower database (hg18). (B) and (G) Box plots showing the methylation levels of samples from each placental group for the DMRs analyzed by bisulphite pyrosequencing. Both DMRs in FAM50B and MCCC1 have higher methylation in digynic than diandric triploid placentas, while they have intermediate methylation in normal placentas and particularly low methylation in CHMs. (C) and (H) Bisulphite cloning and sequencing showing parental origins of methylated and unmethylated alleles (M: maternal alleles; P: paternal alleles). Parental origin was determined by genotyping heterozygous informative SNPs for each sample. The DMRs in both FAM50B and MCCC1 are maternally methylated. Each black circle represents a methylated CpG dinucleotide, and each white circle represents an unmethylated CpG dinucleotide. (D) Quantitative genotyping of methylated alleles by pyrosequencing. SNP rs2239713 is homozygous (GG) in maternal DNA and heterozygous (AG) in foetal (placental) DNA (dispensation order: AAG). Genotyping of the placental sample using a methylation-specific pyrosequencing primer shows a homozygous (GG) pattern, indicating that the DMR associated with the maternally inherited 'G' allele is methylated while the one associated with the paternal 'A' allele is not. (E) and (I) Quantitative genotyping of expressed alleles by pyrosequencing. Both SNPs (E) rs6597007 (dispensation order: GGC) and (I) rs937652 (dispensation order for DNA genotyping: CG; dispensation order for RNA genotyping: CCG) are homozygous in maternal DNA and heterozygous in foetal DNA. Genotyping of cDNA shows a bias towards preferential expression of the paternal alleles. *The pyrosequencing primers used for cDNA genotyping (intron-spanning) in MCCC1 were different from those used for DNA genotyping (Table S10 in Additional file 2), so the peak ratio shown in genotyping the pyrogram of cDNA does not correspond to that for DNA.