Quickly verify claims by accessing the underlying experimental data and figures.
Press Enter β΅ to solve
Fuel Your Discoveries
"The finding of the double helix thus brought us not only joy but great relief. It was unbelievably interesting and immediately allowed us to make a serious proposal for the mechanism of gene duplication."
- James Watson
Quick Explanation
Copied
Paper focus
The study builds deep, base-resolution DNA methylation maps across 18 human tissue types and integrates matched transcriptomes and phased genomes, reporting widespread tissue-specific CG differential methylation, large sets of partially methylated domains (PMDs), and substantial non-CG methylation (mCH) that is linked to tissue programs and gene expression.
Long Explanation
Human body epigenome maps reveal noncanonical DNA methylation variation
Primary outcome: Deep, integrated, base-resolution multi-tissue methylomes reveal tissue-specific CG methylation variation plus widespread adult non-CG methylation (mCH), with correlations to transcription and regulatory features.
1) Visual diagnostics from the paper text (scale/abundance)
The following plots use only the numeric counts explicitly stated in the provided paper text (not additional assumptions).
Number provenance note
The numbers above come verbatim from the provided paper text: 4,073,896 strongly differential CG sites (of 26,474,560 tested; 15.4%), 1,198,132 DMRs, and 719,837 novel DMRs.
2) Visualizing the key biological claims as a causal-logic map
This directed graph separates what the paper measures (data) from what it predicts/infers (models/correlations).
3) What is strongly supported vs what is more correlative
Strongly supported by the paperβs measurements
Tissue structure clusters using both DMR methylation and differentially expressed gene expression, consistent with organ-system-level separation.
Intragenic CG methylation is correlated with expression, with strongest negative correlation not in immediate promoters but downstream up to ~8 kb, as described in the text.
mCH is widespread in differentiated adult tissues with motif-associated sample stratification (TNCAC-like vs NNCAN-like signatures) and an expression correlation that is negative in gene body quantile analyses.
More inference/prediction (not causal proof)
uiDMR functional βclassesβ rely on histone mark profiles and their correlation with expression, and the mapping to regulatory element types is classification-based rather than perturbational causality.
Motif-based methylation prediction (Epigram) demonstrates discriminative predictive ability for tissue-specific methylation from DNA motifs, but predictions are not equivalent to direct mechanistic binding or methylation enzyme recruitment.
XCI escape prediction from gene body mCH is explicitly a predictive/computational result with reported AUC (~0.89 for gene body mCH), but the direction of causality (mCH as driver vs as correlated consequence) is not established by perturbation in the provided text.
4) Uncertainty & critique (skeptical checklist)
Below are key potential blind spots that affect interpretation. Iβm not claiming these flaws exist beyond what the paper text implies; theyβre the standard failure modes to check for in this type of study.
Donor count / tissue heterogeneity
The core tissue mapping uses post-mortem samples from 4 individuals, spanning 18 tissue types.
With small N, some βtissue-specificβ signals could partly reflect donor-specific effects, batch effects, or sampling variationβespecially for mCH where the paper reports motif-based stratification with some tissue replicates disagreeing.
Correlations do not prove causality
The paper links methylation to expression using correlations and distance-to-TSS patterns. That is informative but cannot distinguish whether methylation is upstream regulatory cause vs a consequence of transcriptional state.
Thresholding and DMR definition sensitivity
The definition of βstrongly differentialβ uses a minimum methylation difference threshold (0.3) and DMR aggregation uses sites within 500 bp. These choices can affect counts and downstream functional enrichment.
Functional claims beyond epigenetic marks
For uiDMRs, functional interpretation uses histone marks and motif enrichment, with a classification into enhancer/promoter-like types. This is hypothesis-generating; perturbation validation would be needed to confirm mechanistic roles.
5) Key comparative anchor: mCH and X-inactivation escape prediction
The paper reports a notably high predictive performance for XCI escape status based on gene body mCH. Below is a compact βscore cardβ visualization from the text.
Evidence note
Values are taken from explicit statements in the provided paper text: AUC ~0.79/0.74 for Epigram motif prediction and AUC ~0.89 for gene body mCH predicting XCI escape.
Skeptical interpretation: high AUC supports discriminability, but it does not ensure that mCH is mechanistically causal for XCI escape; it could also reflect correlated chromatin state or allele-specific transcriptional programs.
6) Reproducibility & data access (what you can check)
The paper states that sequencing datasets are available via GEO/SRA, and provides an analyzed-data resource link.
Human tissue methylome/transcriptome datasets: GEO accessions GSE16256, GSE18927, GSE39777, GSE47966.
7) What would most disprove the paperβs central narrative?
Falsification-focused points, grounded in the paperβs own framing:
Replicate the tissue-specific DMR landscape with independent donors and see whether the fraction of strongly differential CG sites and the DMR clustering by organ system remains comparable after controlling for batch and tissue composition.
Test whether non-CG mCH motifs and the reported negative correlation between gene-body mCH and expression persist under independent measurement pipelines and across larger cohorts; address the reported replicate disagreements in some tissues.
Move from prediction to perturbation: show that experimentally altering mCH at the relevant motifs/domains changes gene expression and/or XCI escape status, rather than only improving predictive models.
Author review links (bespoke follow-ups)
Click to ask BGPT for author-specific critical perspectives and likely methodological assumptions.
Feedback:
Updated: April 28, 2026
BGPT Paper Review
Study Novelty
80%
Novelty is high because it integrates base-resolution CG and non-CG methylation across many human tissues with matched transcriptomes and phased genomes, emphasizing mCH in differentiated tissues and linking gene-body mCH to X-inactivation escape prediction; while related epigenome-map work existed, the combined tissue breadth + noncanonical methylation focus is distinctive for its time.
Scientific Quality
80%
Scientific quality is solid: deep sequencing, explicit thresholds, integrated multi-omics, and explicit AUC-style predictive metrics. Main quality caveat is interpretability: many claims rely on correlation/classification from marks and motifs rather than perturbation causality, and the tissue survey uses a small donor number (n=4) which can limit generalization and robustness of allele-specific and non-CG pattern claims.
Study Generality
70%
Findings about widespread tissue-specific methylation and mCH motif stratification likely generalize, but quantitative effect sizes and motif-to-function mappings may be context- and cohort-dependent. The specific predictive framework (e.g., mCH β XCI escape) may transfer, but causality and universality need broader validation.
Study Usefulness
90%
Extremely useful as a reference resource for cis-regulatory element mapping and for generating mechanistic hypotheses about both CG and non-CG methylation variation across human tissues; also provides publicly accessible datasets described in the paper.
Study Reproducibility
80%
Reproducibility is supported by stated GEO/SRA accessions and a public analyzed-data resource. Remaining uncertainty is whether all processing steps/parameters and downstream analytic pipelines are fully available/standardized beyond what is in the paper text.
Explanatory Depth
70%
Depth is moderate: the paper explains patterns (DMRs vs expression distance, uiDMR classes via histone marks, mCH motifs and expression correlations) and proposes regulatory roles, but mechanistic causality remains untested in the described work.
Fetch GEO/SRA methylome and RNA-seq for the stated tissues, compute DMRs with the paperβs thresholds, correlate DMR methylation with nearest-gene expression distance bins, and reproduce the organ-system clustering.
Get emailed when your analysis is done!
We'll email you the results when your analysis is finished.
Hypothesis Graveyard
uiDMRs are not a distinct regulatory class but instead reflect technical/annotation artifacts from DMR calling; if re-calling DMRs with stricter feature-aligned segmentation collapses uiDMR expression correlations, the class would be weakened.
mCH is purely a stochastic byproduct of transcriptional activity and not regulated by sequence motifs; if motif-stratified groups disappear under alternative base-calling/coverage filters, the motif grammar claim would fall.
Quick Explanation
Long Explanation
Top Data Sources
ExportMCP
Keep Exploring
Analysis Wizard
Hypothesis Graveyard
Potential Experiments
Discussion
Fuel Your Discoveries
Ask a Follow-Up
Key Insight
