Research ArticleHUMAN GENETICS

Establishment of environmentally sensitive DNA methylation states in the very early human embryo

See allHide authors and affiliations

Science Advances  11 Jul 2018:
Vol. 4, no. 7, eaat2624
DOI: 10.1126/sciadv.aat2624
  • Fig. 1 Methylation of ME and control regions in samples used for the ME screen.

    Methylation distributions in 687 ME and 5902 control regions. X axes represent methylation in the sample given by the label at the top of the column. Y axes, the same, but given by label at the end of the row. Each point represents mean methylation across all CpGs in a single region. Samples are listed by individual and tissue. HF, hair follicle; PBL, peripheral blood lymphocytes; FT, fat; SB, small bowel. Methylation in MEs is on the bottom (left) of the diagonal; methylation of controls (C) is on the top (right).

  • Fig. 2 Characteristics of ME and control regions.

    (A) Proportion of MEs, control regions, and the whole genome, overlapping predicted chromatin states from three Roadmap samples (one each from endoderm, mesoderm, and ectoderm). Fifteen chromHMM states have been reduced to eight after combining states, which represent similar genomic features. Quiescent loci are those with low signal in all of the histone marks used as input to the chromHMM algorithm; ZNF regions are those overlapping zinc finger protein–encoding genes. (B) Proximity of MEs and controls to ZFP57 binding sites. Black bars show the proportion of regions with a binding site within the distance marked on the x axis. Gray bars show the cumulative proportion of regions within a given distance of a binding site. Red bars show the proportion of regions that overlap a binding site. PBMC, peripheral blood mononuclear cell; TSS, transcription start site.

  • Fig. 3 Methylation at MEs, all-RRBS background, and SoC-associated loci, and potential genetic influences.

    Within each developmental stage in all plots, CpGs are counted once for each replicate for which there was sufficient read depth. (A) Mean methylation at each developmental stage assayed by Guo et al. Solid line: mean methylation in all-RRBS background (n = 3,679,155 CpGs); dashed line: mean methylation at ME regions covered by RRBS (n = 302 regions; 2098 CpGs). (B) Distribution of methylation across developmental stages in all-RRBS background CpGs (left) and at CpGs within ME regions (right). See legend in (C). (C) Distribution of methylation at each stage in previously identified season of conception–associated differentially methylated regions (SoC-DMRs; see Materials and Methods). (D) Proportion of methylation variance explained by genetic variants in cis and in trans at ME-CpGs. These data are obtained from a previous comprehensive screen for methylation quantitative loci (mQTL)–CpGs in a large European study (23).

  • Fig. 4 Methylation dynamics around the gastrulation transition.

    (A) Distribution of CpG methylation in embryonic liver, stratified by methylation level in ICM. Horizontal lines show the median methylation in embryonic liver. Error bars represent 95% bootstrapped confidence intervals. (B) Change in methylation at individual CpGs across the transition. Each line represents the change in methylation at a CpG for a single combination of replicates from each stage. *Because of the large number of data points, “all-RRBS” plot shows a random 0.1% sample of all possible ICM–embryonic liver transitions. (C) Distribution of methylation states in each embryonic liver replicate.

  • Fig. 5 Read-level methylation analysis.

    (A) Distribution of read-level methylation for reads overlapping CpGs with intermediate methylation (10 to 90%) in clustered control regions (left, 31.8K reads) and ME regions (right, 22.5K reads). (B) RHI scores for MEs (n = 133) and control regions (n = 200) covered in the Guo et al. data show a significantly higher average RHI in MEs (Mann-Whitney, P < 5 × 10−5; nine MEs and two control outliers with RHI > 6 not shown). (C) Lollipop plots showing methylation calls overlapping an ME (left, reads ordered by read-level methylation) and a random resampling of the methylation calls, preserving the location of the reads (right). At this ME, the observed methylation calls result in a transition count of 11, while the resampled example has a transition count of 34, which is the median of 1000 resamplings performed on this region. The RHI of this region is thus 34/11 or 3.1.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/7/eaat2624/DC1

    Fig. S1. ME and control region sizes, and Guo et al. methylome coverage.

    Fig. S2. Mean methylation at MEs and clustered control regions assayed by Guo et al.

    Fig. S3. Mean methylation at all CpGs and at MEs and clustered control regions assayed by Zhu et al. (21).

    Fig. S4. Methylation dynamics at the ICM–to–embryonic liver transition.

    Fig. S5. ME background comparisons in other fetal tissues, and methylation in control clusters.

    Table S1. MEs identified in genome-wide screen.

    Table S2. Enrichment of proximal genomic features in MEs.

    Table S3. Number of CpGs covered in each replicate of RRBS data from Guo et al.

    Table S4. Size of ME and control regions, and their coverage in Guo et al. RRBS data.

    Table S5. Overlap of Bak et al. (28) ZFP57-mutant DMRs with MEs.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. ME and control region sizes, and Guo et al. methylome coverage.
    • Fig. S2. Mean methylation at MEs and clustered control regions assayed by Guo et al.
    • Fig. S3. Mean methylation at all CpGs and at MEs and clustered control regions assayed by Zhu et al. (21).
    • Fig. S4. Methylation dynamics at the ICM–to–embryonic liver transition.
    • Fig. S5. ME background comparisons in other fetal tissues, and methylation in control clusters.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Table S1 (Microsoft Excel format). MEs identified in genome-wide screen.
    • Table S2 (Microsoft Excel format). Enrichment of proximal genomic features in MEs.
    • Table S3 (Microsoft Excel format). Number of CpGs covered in each replicate of RRBS data from Guo et al.
    • Table S4 (Microsoft Excel format). Size of ME and control regions, and their coverage in Guo et al. RRBS data.
    • Table S5 (Microsoft Excel format). Overlap of Bak et al. (28) ZFP57-mutant DMRs with MEs.

    Download Tables S1 to S5

    Files in this Data Supplement:

Navigate This Article