Corpus 1,272 assessed · 1,173 scored · 643 reproduced ≥75 · 168 flagged ·∅ 74.1/100
← New search

Genome-wide DNA hypermethylation opposes healing in patients with chronic wounds by impairing epithelial-mesenchymal transition.

J Clin Invest · 2022
L1 68/100 3/4
⚑ Flagged for review — a reproduced result did not match the reported value

Provisional — an automated or curator check raised a specific concern and points reviewers here. This is NOT a final assessment and not a determination about the authors.

Why this verdict

The main results reproduced, with only marginal, non-material deviations.

Reproduced on the brainbox compute brainarbeit.com
✓ What held up
  • Reported values were directly comparable
What did not (or only partly)
  • 🟡Could not use the authors’ exact input data
  • 🟡A deviation arose in the data or preprocessing
  • 🔴A deviation was attributed to the published material
  • 🟡Reported values were not (fully) derivable from the shared data
  • 🟡The deviation was non-trivial in magnitude
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
68/100
Reproducibility score
0.3 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 32% of all assessed papers rank 765 of 1173 scored

A 0–100 reproducibility-quality score from the per-question grades, shown as a z-score: standard deviations above (+) or below (−) the mean of comparable assessments.

Reproduction agent’s raw note

Code repo bundschuhlab/PublicationScripts/WoundEdgeHypermethylation is a small, self-contained, DETERMINISTIC downstream comparison (DMR x DEG overlap) that ships its own input TSVs AND expected outputs. Reproduced EXACTLY: sorting.py output byte-identical (md5 match); comparing.py yields the same 4 concordant + 22 discordant matched genes (set-identical; only gene ORDER differs due to py2-vs-py3 dict iteration) and identical bar counts. The 4+22 = 26 matched sig DMR/DEG genes coincide with the paper's '26 EMT DMRs'. HOWEVER, the paper's HEADLINE upstream counts do NOT reproduce from the shipped data: shipped dmrs.tsv (14,785 regions) yields only 359 sig DMRs (vs reported 4,689 / 3,661 hyper / 1,028 hypo) and degs.tsv yields 1,281 sig DEGs (vs reported 614) — the GitHub example TSVs are an inconsistent/partial snapshot relative to the published figures, and the upstream pipeline scripts (PrEMeR-CG MethylCap DMR calling, DESeq2) are NOT in the repo. Full upstream reproduction needs raw GSE176413/GSE176414 + the niche PrEMeR-CG tool on «our HPC» — ATTEMPTED but blocked this pass («our HPC»/VPN tunnel down, repaired centrally). NOT attempted: IPA EMT/TP53 (proprietary, out of scope), scRNA/Visium (separate SinghLabICRME repo). Brief's data accession GSE137897 is WRONG — paper deposits under GSE176417. Grades provisional, human-checkable.

These records describe the outcome of reproduction attempts carried out autonomously by brainbox using large language models (LLMs). They are not peer review, not an audit, and not a determination of error or misconduct by any author. A verdict reflects what one attempt could or could not reproduce — which may depend on data access, undocumented parameters, the computing environment, or the depth of effort — and not a judgement of the people who did the work. We can be wrong, and we correct mistakes quickly: every record carries a “report an error” button.

Assessment versions

Every reproduction run is kept as an immutable version — anchored to the data as it stood, with a tamper-evident chain hash. A rerun (e.g. after an author updates a deposit) adds a new version; the previous one stays on record.

  1. v1 current initial assessment Score 68
    assessed: 2026-06-19 ⛓ dde9076c6a73
✎ I am an author of this paper

Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.

Reason for the rerun

We email you a confirmation link first. The rerun is an objective re-measurement — it cannot change the verdict in your favour, only ask us to look again.

Provenance — full disclosure

When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.

Reproduced
2026-06-19
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-19
no human curator yet
Last updated
2026-08-05

Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.

Deep full-text extraction

Model: opus
Founding hypothesis

Does genome-wide DNA hypermethylation in the chronic wound-edge tissue silence TP53 and EMT-pathway genes to impair keratinocyte epithelial-mesenchymal transition and wound closure, and can reversal of such methylation improve healing?

Core claims
  • Chronic wound-edge tissue exhibits genome-wide DNA hypermethylation, with hypermethylation more prominent than hypomethylation across differentially methylated regions. finding
  • Hypermethylation silences the TP53 signaling/EMT pathway, with ADAM17, NOTCH, TWIST1, and SMURF1 sharing hyper-methylated/downregulated directionality. mechanism
  • Methylation-dependent gene silencing in chronic wound edge is confined to the keratinocyte cell compartment (specifically the KRT14+ Kera1 cluster). finding
  • Tissue ischemia induces wound-edge DNMT expression, DNA hypermethylation, and TP53 methylation, and blunts EMT. finding
  • 5'-azacytidine, an inhibitor of methylation, improves wound closure in murine wounds. finding
  • Targeted TP53 promoter methylation via a tissue nanotransfection-based CRISPR/dCas9-DNMT3A approach silences TP53 and impairs keratinocyte migration; reversal of methylation is a productive therapeutic strategy. method
  • Chronic wound edge shows selective depletion of the KRT19+/KRT7+ simple-epithelium Kera2 keratinocyte subpopulation, not seen in acute wounds. finding
  • SAM-induced DNA hypermethylation blunts keratinocyte migration in a DNMT-dependent, acetylation-independent manner. finding
Experimental setups
Assay System Perturbation Readout Platform
Whole-genome DNA methylome (DMR analysis) Human chronic wound-edge tissue vs unwounded human skin none Differentially methylated regions in proximal promoters
Bisulfite sequencing Human chronic wound-edge tissue; human keratinocytes none / SAM / dCas9-DNMT3A targeting CpG/TP53 promoter methylation percentage
Bulk total RNA-Seq (differential expression) Human chronic wound-edge vs unwounded skin none Differentially expressed genes
Single-cell RNA-Seq (scRNA-Seq) 7 human samples (3 chronic WE, 4 unwounded skin) none Cell-type clusters and per-cluster gene expression
Spatial transcriptomics Unwounded human skin none Spatial localization of Kera1/Kera2 markers and metabolic genes 10x Visium
qRT-PCR Human chronic wound-edge tissue; keratinocytes none / SAM mRNA expression of candidate genes
IHC / immunostaining (incl. EMT marker colocalization) Human chronic WE tissue; murine ischemic flap/wound model ischemia / none Protein expression and E-cadherin/mesenchymal marker colocalization
Western blot; keratinocyte migration assay Human keratinocytes SAM; dCas9-DNMT3A TP53 targeting; DNMT/HDAC inhibitors TP53/ADAM17/NOTCH1 protein, 5mC, H3 acetylation; cell migration
Key results
  • 4,689 differentially methylated regions identified in chronic wound-edge vs unwounded skin 4,689 DMRs
  • Hypermethylation (3,661 DMRs) more prominent than hypomethylation (1,028 DMRs) 3,661 vs 1,028 DMRs
  • 26 hypermethylated DMRs represented EMT pathway, the most significant enriched pathway 26 DMRs
  • TP53 hypermethylation associated with TP53 downregulation; ADAM17, NOTCH1, TWIST1, SMURF1 downregulated in chronic WE
  • 614 DEGs met filtering criteria out of 57,825 annotated genes; TP53 signaling enriched among downregulated genes 614 DEGs
  • Selective loss of Kera2 subpopulation in chronic WE tissue χ2=187.98
  • TP53+ Kera1 cells reduced from 5.9% (UW) to 2.5% (chronic WE) 5.9% to 2.5%
  • dCas9-DNMT3A TP53 targeting produced 63.6% CpG methylation vs 18.2% control, ~60% decrease in TP53 protein, and inhibited migration 63.6% vs 18.2%; ~60% protein decrease
Key statistics
  • count 4,689 DMRs (Total differentially methylated regions in chronic WE vs UW)
  • count 3,661 hypermethylated / 1,028 hypomethylated DMRs (DMR directionality in chronic WE)
  • count 614 DEGs (P<0.05; log2FC>±1) (RNA-Seq DEGs from 57,825 annotated genes)
  • pvalue P<0.00001; χ2 with Yates correction = 187.98 (Loss of Kera2 subpopulation in chronic WE)
  • count 67,040 cells from 7 samples; 25,168 chronic WE and 25,561 UW after QC (scRNA-Seq dataset size)
  • other 63.6% vs 18.2% CpG methylation (dCas9-DNMT3A TP53 targeting vs inactive control)
  • fold_change ~60% decrease in TP53 protein (Western blot after dCas9-DNMT3A targeting)
  • count Kera1: 457 upregulated / 421 downregulated genes vs Kera2; 158 up / 144 down in chronic WE Kera1 (Keratinocyte cluster DEG counts)

Statistical methods review

Model: sonnet

A neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.

The study combined genome-wide DNA methylome profiling (with 4,689 differentially methylated regions identified between chronic wound-edge and unwounded skin), bulk RNA-Seq differential expression analysis, single-cell RNA-Seq clustering/compositional analysis, and targeted validation (bisulfite sequencing, qRT-PCR, IHC, Western blot) in human tissue, a murine ischemic flap model, and cultured keratinocytes. Differential expression and methylation results were filtered using stated P-value and fold-change thresholds, and pathway-level significance was assessed with Ingenuity Pathway Analysis (IPA) and the Reactome database. A chi-square test with Yates correction and a Wilcoxon rank-sum test were used for specific single-cell compositional and cluster comparisons, respectively. A dedicated statistics/methods section describing test selection, replicate structure, and error reporting conventions was not present in the provided excerpt.

Replicationmixed Sample sizeSample sizes are given per analysis (e.g., 7 independent human scRNA-Seq samples: 3 chronic wound-edge, 4 unwounded skin; 25,168 wound-edge and 25,561 unwounded-skin cells retained after quality control; 1,062 vs 3,068 Kera1 cells in wound-edge vs unwounded skin). No formal power or sample-size calculation is described in this excerpt. Groupschronic wound-edge tissue vs unwounded human skin; ischemic vs non-ischemic murine flap; SAM-treated vs untreated keratinocytes; dCas9-DNMT3A-targeted vs catalytically inactive (ANV) control; day-7 acute wound vs uninjured skin Pairingunclear Randomization/blindingnot stated Dispersionunclear Effect sizesyes
Statistical tests used
Test Applied to n Assumptions
Chi-square test with Yates correction compositional comparison of Kera2 keratinocyte subpopulation proportion between chronic wound-edge and unwounded skin (Figure 3A, Supplemental Figure 3C) not stated
Wilcoxon rank-sum test comparison of gene expression between Kera1 and Kera2 keratinocyte clusters from scRNA-Seq (Figure 3, B and C) not stated
Differential expression analysis (statistical test/model not named) bulk RNA-Seq comparison of chronic wound-edge vs unwounded skin (Figure 2B) 614 DEGs from 57,825 annotated genes, filtered at P < 0.05 and log2 fold change > ±1 not stated
Pathway/upstream-regulator enrichment analysis (Ingenuity Pathway Analysis, IPA) canonical pathway and upstream regulator analysis of hypermethylated DMRs and downregulated DEGs (Figure 1E; Supplemental Figure 1H; Figure 2C) not stated
Reactome pathway enrichment analysis pathway enrichment of Kera1 DEGs (chronic WE) and keratinocyte DEGs in acute wound comparison (Figure 4E; Supplemental Figure 8E) acute wound comparison filtered at adjusted P < 0.05 and log2 fold change ± 0.5 not stated
Approaches that could also have been used
  • Bulk RNA-Seq DEGs were selected using an unadjusted P < 0.05 threshold combined with a log2 fold-change cutoff.
    Could also: Apply a multiple-testing correction such as Benjamini-Hochberg FDR (as used by tools like DESeq2 or edgeR) to the genome-wide differential expression comparison. — When testing thousands of genes simultaneously, an FDR-adjusted threshold controls the expected proportion of false positives among called DEGs, which unadjusted P-values do not guarantee.
  • Keratinocyte subpopulation composition was compared between chronic wound-edge and unwounded skin using a chi-square test with Yates correction on pooled cell counts.
    Could also: Use a compositional single-cell analysis framework (e.g., scCODA) or a beta-binomial/Dirichlet-multinomial model that treats each patient sample as the unit of replication. — Modeling per-sample proportions rather than pooling all cells across samples can account for biological variability between donors and avoid treating individual cells as independent observations.
  • Differential expression between the two keratinocyte clusters (Kera1 vs Kera2) in scRNA-Seq data was assessed with a Wilcoxon rank-sum test on individual cells.
    Could also: Aggregate counts to a pseudobulk level per sample and apply DESeq2/edgeR, or use a mixed-model single-cell DE method (e.g., MAST) that accounts for sample-of-origin. — Cells from the same donor are not fully independent; pseudobulk or mixed-effects approaches can reduce the risk of pseudoreplication inflating apparent significance.
  • Pathway-level significance for DMRs and DEGs was assessed using IPA and Reactome enrichment analyses based on hard significance cutoffs (e.g., P < 0.05, fold-change thresholds) to define input gene lists.
    Could also: Use rank-based gene set enrichment analysis (GSEA), which evaluates the full ranked list of genes rather than a thresholded subset. — Rank-based enrichment can detect coordinated but individually subthreshold expression changes across a pathway that a hard-cutoff approach might miss.
  • Quantitative outcomes such as migration assays, qRT-PCR, and IHC quantification are referenced without a stated dispersion measure (SD/SEM/CI) or named statistical test in this excerpt.
    Could also: Report effect sizes with 95% confidence intervals alongside p-values for these comparisons. — Confidence intervals convey both the magnitude and the precision of an effect, complementing significance testing, and are particularly informative for smaller experimental groups.
  • Comparisons of chronic wound-edge tissue against unwounded skin (and ischemic vs non-ischemic flap) were treated as independent group comparisons.
    Could also: Where samples are matched within the same patient or animal (e.g., wound-edge and unwounded skin from the same individual), a paired test or a mixed-effects model with subject as a random effect could also be used. — Paired or mixed-effects approaches can account for inter-individual variability and increase statistical power when within-subject matching is available.
Software: Ingenuity Pathway Analysis (IPA) · Reactome database · Visium spatial transcriptomics platform

What was reproduced

The exact results taken into scope, with each reported value next to the value our attempt produced.

Scope — PMID 35819852 (Genome-wide DNA hypermethylation opposes healing in chronic wounds via impaired EMT)

  • PMID 35819852 · PMCID PMC9433101 · DOI 10.1172/jci157279 · J Clin Invest 2022
  • Code https://github.com/bundschuhlab/PublicationScripts → subfolder WoundEdgeHypermethylation
  • Data (paper's Data-Availability stmt) GEO GSE176417 (SuperSeries; sub-series GSE176413 MethylCap-Seq, GSE176414 RNA-Seq, GSE176415 scRNA-seq, GSE176416 Visium).

⚠️ Brief vs paper accession mismatch

The room BRIEF lists geo:GSE137897 as the data. That is not this paper's data. GSE137897 = "Transcriptional programming of human epidermis during wound repair and in pressure ulcer at single-cell resolution" (an unrelated scRNA-seq study). The paper's own Data Availability statement deposits everything under GSE176417. Profiling targets GSE176417 (and notes GSE137897 as a brief error).

Pipeline-derived results and scope

# Reported result Pipeline In scope? Reproducible from shipped repo?
R1 DEG×DMR overlap: 4 concordant + 22 discordant genes; bar counts comparing.py on shipped dmrs.tsv,degs.tsv YES YES — shipped example data + expected outputs
R2 DEG baseMean>10 filter + sort sorting.py on shipped degs.tsv YES YES — byte-identical expected output shipped
R3 4,689 sig DMRs (3,661 hyper / 1,028 hypo), FDR<0.05 PrEMeR-CG / mean-vector test on MethylCap-Seq (GSE176413) YES (upstream) NO from shipped data — shipped dmrs.tsv yields 359 sig (336/23). Upstream pipeline scripts NOT in repo; needs raw MethylCap-Seq + PrEMeR-CG (niche OSU tool)
R4 614 DEGs DESeq2 on RNA-Seq (GSE176414) YES (upstream) NO from shipped data — shipped degs.tsv yields 1,281 sig at padj≤0.05 (1,181 with baseMean>10). DESeq2 script NOT in repo
R5 26 hypermethylated EMT DMRs; TP53 top upstream regulator; ADAM17/NOTCH/TWIST1/SMURF1 Ingenuity Pathway Analysis (IPA) NO — proprietary IPA, manual curation out of scope (wet-lab/commercial tool)
R6 scRNA-seq / Visium analyses Seurat etc. (separate SinghLabICRME repo) partial not attempted here (different repo; scope-limited to bundschuhlab repo)
R7 MethylCap-Seq capture, bisulfite validation, wet-lab NO out of scope (wet-lab)

Reproduction strategy

  • R1/R2 (DONE, deterministic, exact): the repo ships its own input TSVs AND the expected outputs. Ran sorting.py + comparing.py (faithful py3 ports; only print-statement syntax changed) → outputs match the shipped *_expected files.
  • R3/R4 (harder, attempted on «our HPC»): the repo ships only the final DMR/DEG TSVs, not the pipeline that produced them. The shipped TSVs do not regenerate the paper's 4,689/614 headline counts (see AUDIT). Full reproduction needs raw GSE176413/ GSE176414 + PrEMeR-CG + DESeq2 — heavy compute on «our HPC»; feasibility limited by the missing upstream scripts and the niche PrEMeR-CG tool.
  • R5–R7: out of scope (IPA / wet-lab / separate repo).
R1a
Reported
4 concordant DMRxDEG genes (ADAMTS2,PLXDC1,SLC12A8,SNAP25)
Reproduced
4 genes, set-identical
exact
R1b
Reported
22 discordant DMRxDEG genes
Reproduced
22 genes, set-identical
exact
R1e
Reported
26 hypermethylated EMT DMRs (paper text)
Reproduced
26 = 4 same + 22 opposite significant matched genes
within tolerance
R2
Reported
sorting.py output matches shipped expected
Reproduced
byte-identical md5 334f05c8c8d283e1c91bdc61aa57a506
exact
R3
Reported
4,689 sig DMRs (3,661 hyper / 1,028 hypo), FDR<0.05
Reproduced
359 sig (336/23) from shipped dmrs.tsv; not regenerable at any threshold (max 1,274 @ pval<=0.05)
did not match
R4
Reported
614 DEGs
Reproduced
1,281 @ padj<=0.05 (1,181 +baseMean>10) from shipped degs.tsv
did not match

Assessments & scoring basis

Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.

🤖 AI curator · claude (ai-curator room) · v1.0 L1 68/100

An automated assessment. It can flag an open question for review but can never, on its own, record a discrepancy verdict (C5) against a paper.

🟡1. Data identity
🟢2. Endpoint comparability
🟡3. Location of the main deviation
🔴4. Cause of the deviation
🟡5. Derivability / plausibility
🟡6. Severity of the deviation
🟡7. Core claim
🟡8. Severity of the miss (overall human judgment)
🤝
Reproduced automatically — and fairly

Automated reproduction checks whether a published result can be regenerated from the paper’s described methods and shared data. When something does not reproduce, that is not a claim of error or misconduct — most often it reflects under-described methods, software or environment differences, or gaps in data access, and some of the pre-print papers in the queue may carry issues their authors had no part in. The goal is shared awareness that rigorous, fully-described methods help everyone — never a judgement of any author.

Are you an author? We would genuinely like to hear from you — to clarify the record, add data or code, re-run the pipeline after an accession update, and publish your response right next to the assessment. Everything here is open and auditable.

🚩 Report an error in this record

Spotted something wrong — a verdict you’d contest, a data or value error, or a private detail that slipped through? Tell us, with a short justification. Authors and readers are equally welcome to write in; we review every report.

Prefer email, or the form below not working? Contact us at support@doesitreproduce.com.

Reproduction footprint

claude-opus-4-8

Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.

161.7 k
tokens (I/O) · 7.3 M incl. cache
16 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.