Journal: Nature Communications

Article Title: Atlas of imprinted and allele-specific DNA methylation in the human body

doi: 10.1038/s41467-025-57433-1

Figure Lengend Snippet: Using fragment-level analysis, we identified 324,759 genomic loci showing a mixture of fully methylated and fully unmethylated DNA fragments. Genetic variation at neighboring SNPs was used to split the fragments by genotype, and to identify 34,426 loci that show allele-specific methylation. These were analyzed across multiple donors and classified as sequence-dependent allele-specific methylation (SD-ASMs, bottom-left) where methylation consistently segregates with one allele, or parental allele-specific methylation (bottom-right) if both methylated and unmethylated epialleles exist, regardless of the genotype, suggesting that methylation is associated with parental, rather than genetic, origin. Overall, we identified 460 parental-ASM loci, including most known imprinted DMRs, as well as multiple novel regions. 78 of these regions are associated with imprinted genes. One such novel region within CHD7 , a gene associated with CHARGE syndrome, was validated as maternally methylated. Remarkably, some of these regions also show cell-type-specific effects, including escape of allele-specific methylation.

Article Snippet: To a large extent, this is due to the fact that most genome-wide methylome datasets are based on DNA methylation arrays (Illumina BeadChip 450 K and EPIC), and are limited to a predefined set of CpGs, capturing only 1.5%-3% of the 28 million methylation sites in the human genome.

Techniques: Methylation, Sequencing

Journal: Nature Communications

Article Title: Atlas of imprinted and allele-specific DNA methylation in the human body

doi: 10.1038/s41467-025-57433-1

Figure Lengend Snippet: A A computational algorithm identifies bimodal regions ( n = 324,759), by analyzing deeply sequenced methylomes from 39 cell types. Shown is a bimodal region (chr19:54039871-54043130, hg19, highlighted) where 51% of DNA fragments are methylated (black), and 46% are unmethylated, in a colon macrophage sample purified from a single donor. B Similarly, DNA fragments from Adipocytes were split by a common T/G SNP (rs2071094, chr11:2021164) to show allele-specific methylation. Fragments carrying the T allele are unmethylated (white), whereas G allele fragments are methylated. ( C ) Contingency table of alleles by methylation, as shown in (B). All 29 unmethylated fragments are from the G genotype, whereas all 22 methylated ones carry the T genotype (adj. p 7.1E-19 ≤ 7.1E-19, Fisher’s exact). D Genetic/Epigenetic table across multiple samples/cell types (rs9330298, chr1:153590254, hg19). Here, for all samples (homozygous or heterozygous), unmethylated fragments (U) have the G genotype, whereas the methylated fragments (M) are associated with the alternative T genotype, consistent with sequence-dependent allele-specific methylation (SD-ASM). E A similar table for rs80269905 (chr11:2720873), is consistent with parental allele-specific methylation. All samples are bimodal (showing both U and M fragments), and heterozygous samples are associated with allele-specific bimodal patterns but switch across different donors.

Article Snippet: To a large extent, this is due to the fact that most genome-wide methylome datasets are based on DNA methylation arrays (Illumina BeadChip 450 K and EPIC), and are limited to a predefined set of CpGs, capturing only 1.5%-3% of the 28 million methylation sites in the human genome.

Techniques: Methylation, Purification, Sequencing

Journal: Nature Communications

Article Title: Atlas of imprinted and allele-specific DNA methylation in the human body

doi: 10.1038/s41467-025-57433-1

Figure Lengend Snippet: A DNA fragments from adipocytes purified from a single donor, at the known imprinted differentially methylated region (iDMR) of the IGF2 gene (chr11:2018812-2024740, hg19). 52% of fragments are fully methylated (black circles) whereas 45% are fully unmethylated (white circles). B Stacked bars showing the percent of methylated (red), unmethylated (green) or mixed (yellow) DNA fragments, across 202 purified samples, spanning from 39 cell types, where a nearly balanced pattern of 1:1 ratio between unmethylated and methylated fragments is shown for this known imprinted region. C Same for the known iDMR of IGF2R (chr6:160426558-160427561, hg19), showing a bimodal (imprinted) pattern across all samples, except for small intestine and colon epithelial cells ( n = 13 samples), where both alleles are fully methylated (right, bottom). Grey asterisks mark samples from a single donor who exhibits bi-allelic methylation.

Article Snippet: To a large extent, this is due to the fact that most genome-wide methylome datasets are based on DNA methylation arrays (Illumina BeadChip 450 K and EPIC), and are limited to a predefined set of CpGs, capturing only 1.5%-3% of the 28 million methylation sites in the human genome.

Techniques: Purification, Methylation

Journal: Nature Communications

Article Title: Atlas of imprinted and allele-specific DNA methylation in the human body

doi: 10.1038/s41467-025-57433-1

Figure Lengend Snippet: A Average CpG methylation plot (chr11:2018807-2021899) within the known iDMR for H19 / IGF2 . The highlighted 692 bp patch (chr11:2020097-2020789) shows monoallelic methylation in purified granulocytes and T cells samples (as expected) but is fully methylated in hepatocytes, and pancreas ductal epithelial samples. An average of 50% methylation is expected for known iDMRs (dotted line). B Sequenced DNA fragments from within the highlighted region, revealing intra-iDMR biallelic methylation. Indeed, almost all fragments from the hepatocytes and pancreas are fully methylated, compared to half of granulocyte and T cell DNA fragments.

Article Snippet: To a large extent, this is due to the fact that most genome-wide methylome datasets are based on DNA methylation arrays (Illumina BeadChip 450 K and EPIC), and are limited to a predefined set of CpGs, capturing only 1.5%-3% of the 28 million methylation sites in the human genome.

Techniques: CpG Methylation Assay, Methylation, Purification

Journal: Nature Communications

Article Title: Atlas of imprinted and allele-specific DNA methylation in the human body

doi: 10.1038/s41467-025-57433-1

Figure Lengend Snippet: A Fragment-level analysis of a novel parent-of-origin cell-type-specific ASM region we identified (chr8:61627212-61627412, hg19) shows biallelic methylation (red) in endothelial cells, neurons, fibroblasts, and blood cells, but bimodal methylation patterns in hepatocytes and epithelial cells (1:1 ratio of fully methylated and fully unmethylated sequenced fragments of ≥3 CpGs). B – F Genetic/epigenetic analysis of parental allele-specific methylation in tongue-epithelial cells, validated across 15 families (a total of 33 children and their parents, Data ). For each trio, we used targeted-PCR next-generation sequencing (after bisulfite conversion) to measure the genotype (rs7826035, C/T, chr8:61627312 hg19) and the methylation status of six CpG sites (chr8:61627190-61627349 hg19, measured on the bottom strand). B A trio (family ID IMP017) showing homozygous C/C for the mother, with C/T heterozygosity for the father. The child T allele is therefore of paternal origin. Blue bars correspond to relative allelic read count. Unmethylated DNA fragments (green bars) are limited to the parental T allele, suggesting maternal-specific methylation. C same as ( B ) for a family with three heterozygous children (family ID IMP012). D A family where the T allele of heterozygous child 1 is maternal, whereas unmethylated fragments (green) are all from the paternal C allele. Two additional siblings are C/C homozygous and not shown (family ID IMP005). E – F Examples of a C/C homozygous family and a C/T heterozygous family, where the parent-of-origin of unmethylated fragments cannot be associated with a parent-of-origin (family IDs IMP014, IMP011). All remaining families were homozygous (C/C, not shown). G Bimodal tissue count track (orange) and -log10 of Fisher’s method for combining Fisher’s exact p values for allele-specific methylation at individual samples (blue track), along with a gene annotation track showing the region surrounding the identified DMR.

Article Snippet: To a large extent, this is due to the fact that most genome-wide methylome datasets are based on DNA methylation arrays (Illumina BeadChip 450 K and EPIC), and are limited to a predefined set of CpGs, capturing only 1.5%-3% of the 28 million methylation sites in the human genome.

Techniques: Methylation, Next-Generation Sequencing