Copied to clipboard!

Mitigating genotype-driven bias improves atlas-scale single-cell DNA methylome profiling

Single-cell DNA methylation studies often include cells from individuals with distinct genotypes, where C-to-T variants are misread as bisulfite-converted unmethylated cytosines, conflating genotype with DNA methylation differences. We mitigate this variant-driven bias by masking variant positions before DNA methylation quantification and feature selection. Using sciMET profiles of blood nuclei from eight healthy donors with diverse genetic backgrounds, we show that masking reduces donor-associated confounding in clustering and spurious sample-specific differentially methylated regions within the same cell type. This provides practical solutions for improving single-cell DNA methylation analyses in atlas-scale, pooled experimental designs.

Type: Epigenetics
Archive: European Genome-phenome Archive (EGA)

5 Datasets

Click on a Dataset ID in the table below to learn more, and to find out who to contact about access to these data

Dataset ID	Description	Technology	Samples
EGAD50000003173	Single-cell DNA methylation profiling (sciMET) of buffy coat nuclei from eight healthy adult donors (3 male, 5 female; seven European and one Asian ancestry), to characterize the effects of donor genetic diversity in pooled single-cell DNA methylation experiments. The dataset was generated using a pooled design: all eight donors mixed before tagmentation, 5,000 nuclei per well, donors assigned afterwards with cellsnp-lite and Vireo). The libraries followed the sciMET v2 protocol instead of the commercial kit, with modified ligation (T4 PNK/T4 ligase with PEG 8000, 1,3-propanediol, ETSSB and H10 P5 adapter) and amplification (Kapa HiFi Uracil, 15 cycles) reactions.	Illumina NovaSeq X	1
EGAD50000003174	Single-cell DNA methylation profiling (sciMET) of buffy coat nuclei from eight healthy adult donors (3 male, 5 female; seven European and one Asian ancestry), to characterize the effects of donor genetic diversity in pooled single-cell DNA methylation experiments. Nuclei were fixed, nucleosome-depleted and tagmented with the Single-cell DNA Methylation Kit (v1.1, Scale Biosciences/10x Genomics), with two modifications: HEPES resuspension buffer was used instead of the kit’s Met Wash buffer, and centrifugation steps during nuclei preparation and tagmentation were kept at 300 × g. Each donor was kept separate on the tagmentation plates (45 wells per donor, 5,000 nuclei per well) and received a donor-specific set of tagmentation indices, so donor identity could be assigned from the barcodes. Nuclei were stained with DRAQ5 and singlets were sorted by FACS. Libraries were prepared with the v1.1 kit. This dataset (4,240 nuclei) provided ground-truth donor identities for pseudobulk genotype calling (Revelio double-masking, then FreeBayes), and those genotypes were used downstream to demultiplex the pooled datasets.	Illumina NovaSeq X	1
EGAD50000003175	Single-cell DNA methylation profiling (sciMET) of buffy coat nuclei from eight healthy adult donors (3 male, 5 female; seven European and one Asian ancestry), to characterize the effects of donor genetic diversity in pooled single-cell DNA methylation experiments. The dataset was generated using a pooled design: all eight donors mixed before tagmentation, 5,000 nuclei per well, donors assigned afterwards with cellsnp-lite and Vireo). The libraries followed the sciMET v2 protocol instead of the commercial kit, with modified ligation (T4 PNK/T4 ligase with PEG 8000, 1,3-propanediol, ETSSB and H10 P5 adapter) and amplification (Kapa HiFi Uracil, 15 cycles) reactions.	Illumina NovaSeq X	1
EGAD50000003176	Single-cell DNA methylation profiling (sciMET) of buffy coat nuclei from eight healthy adult donors (3 male, 5 female; seven European and one Asian ancestry), to characterize the effects of donor genetic diversity in pooled single-cell DNA methylation experiments. The dataset was generated using a pooled design: all eight donors mixed before tagmentation, 5,000 nuclei per well, donors assigned afterwards with cellsnp-lite and Vireo). The libraries followed the sciMET v2 protocol instead of the commercial kit, with modified ligation (T4 PNK/T4 ligase with PEG 8000, 1,3-propanediol, ETSSB and H10 P5 adapter) and amplification (Kapa HiFi Uracil, 15 cycles) reactions.	Illumina NovaSeq X	1
EGAD50000003177	Single-cell DNA methylation profiling (sciMET) of buffy coat nuclei from eight healthy adult donors (3 male, 5 female; seven European and one Asian ancestry), to characterize the effects of donor genetic diversity in pooled single-cell DNA methylation experiments. Nuclei were isolated and processed as in dataset 1, but nuclei from all donors were pooled before loading into the tagmentation plates (5,000 nuclei per well), so barcodes were not linked to donors in advance. Libraries were prepared with the Single-cell DNA Methylation Kit (v1.1). Donor identity was assigned afterwards from genetic variants using cellsnp-lite and Vireo, with the previously called donor genotypes, which also flagged heterotypic doublets.	Illumina NovaSeq X	1