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

Precise modulation of BRG1 levels reveals features of mSWI/SNF dosage sensitivity.

Nat Genet · 2025
L1 75/100 PQI 92
Why this verdict

The main results reproduced: recomputed values matched the published ones within tolerance.

Reproduced on the brainbox compute brainarbeit.com
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q5 · Derivability / plausibility 🟡
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +6
✓ What held up
  • No relevant deviation in data/preprocessing
  • Any deviation was negligible
What did not (or only partly)
  • 🟡Could not use the authors’ exact input data
  • 🟡Reported values were only indirectly comparable
  • 🟡A deviation was attributed to the published material
  • 🟡Reported values were not (fully) derivable from the shared data
  • 🟡The central claim did not (fully) hold under reproduction
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
75/100
Reproducibility score
at the mean
vs. all fields · 1173 studies
🎯 Scores higher than 45% of all assessed papers rank 612 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

Described well enough for the resource, NOT 1:1 for the figure. The brief's 'Code' link (guifengwei/ChromHMM_mESC_mm10) is a THIRD-PARTY precomputed 12-state ChromHMM mESC mm10 chromatin-state map that the paper downloaded (the paper's own code is YiZhang-lab/Brg1_dTag; its own data = GSE274469+GSE294015; the brief's GSE30206 is a text-mining mismatch). I reproduced, deterministically on «our HPC»/«infra», the properties of that shipped map and confirmed they match the paper's description EXACTLY: 12 states (C1) and all 6 functional categories (C2); the dense annotation is internally consistent with the segmentation (C3, dense bp == segments bp). I also reproduced the paper's described enrichment METHOD (Fisher vs bedtools shuffleBed background) on the shipped map using public mm10 TSS, getting the biologically-expected result (ActivePromoter 63.7x enriched, p~0) which confirms the map's promoter labelling is meaningful (C5). I did NOT attempt the paper's exact Fig 4d/4e chromatin-state enrichment of the BRG1-dependent enhancer groups G1-G5: those enhancer-group coordinates are not deposited (GEO ships only per-sample bigwigs; the authors' R scripts read local-filesystem intermediates), and Fig 4d has no reported source-data numbers to compare against, so it is not 1:1 reproducible without rebuilding the entire upstream ATAC->DAR->clustering pipeline (the heavy 80%). No fabrication concern for the reproduced items; Fig-4d remains an auditability gap, not evidence of fabrication. All grades provisional, for human audit.

💻 Code ↗ 🗄 Data: GSE30206

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 75
    assessed: 2026-06-14 ⛓ 05496b1cbd65
✎ 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
2026-06-14
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-15
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

How does the dosage of the mSWI/SNF complex affect its function? Using precise control of BRG1 (the ATPase subunit) protein levels, the paper tests how BRG1 dosage governs chromatin binding, accessibility and transcription to explain dosage-sensitive effects of mSWI/SNF defects.

Core claims
  • BRG1 chromatin binding decreases linearly and proportionally with BRG1 protein dosage, independent of TFs or histone modifications (92.2% of binding peaks follow a linear model). finding
  • Chromatin accessibility shows genomic-feature-specific BRG1 dependence: enhancers (44.2%) depend on BRG1 far more than promoters (16.1%) or insulators (17.8%). finding
  • About half of BRG1-dependent enhancers, and especially superenhancers, exhibit a buffered (nonlinear) response to BRG1 loss, while weak enhancers are more sensitive. finding
  • At BRG1-independent promoters and insulators, NFYA and the SNF2H–CTCF axis respectively maintain chromatin accessibility, acting as compensatory mechanisms. mechanism
  • Transcription exhibits a predominantly buffered response to changes in BRG1 levels, with downregulated genes being predominantly buffered genes. finding
  • A dTAG degron knock-in at the Brg1 locus enables precise, dose-dependent control of BRG1 protein levels (degradation complete within 60 min) without disrupting mSWI/SNF complex assembly. method
  • BRG1-dosage-sensitive accessibility regulation and its role in MYC expression are conserved in human lung epithelial cells. finding
  • BRG1 depletion triggers increased PRC1/PRC2 (RING1B/SUZ12) binding at BRG1-dependent promoters, indicating broader chromatin reorganization. mechanism
Experimental setups
Assay System Perturbation Readout Platform
Western blot Brg1-FKBP12-F36V knock-in and wild-type mES cells dTAG13-induced BRG1 degradation (dose titration) relative BRG1 protein abundance (normalized to tubulin)
CUT&RUN Brg1-knock-in mES cells dTAG13 at five concentrations (BRG1 dosage modulation) genome-wide BRG1 binding peaks/signal at promoters, enhancers, insulators
ATAC-seq Brg1-knock-in mES cells dTAG13 BRG1 depletion (dose titration) chromatin accessibility / differentially accessible regions at regulatory elements
Co-immunoprecipitation Brg1-knock-in vs untagged wild-type mES cells degron tag (none/AU-15330 context) BRG1 complex formation with ARID1A, PBRM1, BRD9
Immunostaining Brg1-knock-in mES cells dTAG13 treatment population-wide BRG1 protein levels
ATAC-seq (dual degradation) Snf2h-FKBP12-F36V knock-in mES cells SNF2H degradation by dTAG13 and/or BRG1/BRM/PBRM1 degradation by AU-15330 chromatin accessibility at BRG1-dependent vs independent insulators
ChIP-seq reanalysis (public datasets) mES cells none (NFYA/CTCF/OCT4 occupancy); Nfya knockdown reanalysis NFYA, CTCF, OCT4 binding and accessibility at BRG1-dependent/independent elements
ATAC-seq / accessibility profiling human lung epithelial cells BRG1 modulation conserved BRG1-dosage-sensitive accessibility and MYC expression
Key results
  • 92.2% of all 20,326 BRG1-binding peaks followed the linear response model 92.2% of 20,326 sites
  • BRG1 binding peak numbers dropped by ~half at 0.3 nM dTAG13 (78% protein remaining) and were nearly abolished at 10 nM (12% protein remaining) ~50% at 78% protein; near-zero at 12% protein
  • 44.2% of BRG1-bound enhancers lost accessibility versus 16.1% of promoters and 17.8% of insulators 44.2% vs 16.1% vs 17.8%
  • 83.2% of all differentially accessible regions were located at enhancers (11.1% promoters, 5.7% insulators) 83.2% / 11.1% / 5.7%
  • BRG1-dependent regulatory elements split approximately evenly between linear (50.7%) and buffered (49.3%) responses; promoters and insulators preferred buffered, enhancers slightly preferred linear 50.7% linear vs 49.3% buffered
  • BRG1 binding was enriched at promoters (23.9%), enhancers (65.0%) and insulators (11.1%) 23.9% / 65.0% / 11.1%
  • Combined SNF2H+BRG1 depletion further reduced accessibility at BRG1-dependent insulators but not independent insulators; SNF2H loss alone reduced accessibility most at BRG1-independent insulators
  • NFYA and CTCF preferentially bound BRG1-independent promoters and insulators respectively, and accessibility at NFYA/CTCF sites remained unchanged after BRG1 depletion
Key statistics
  • count 20,326 BRG1-binding sites (total BRG1-binding sites analyzed for delta values)
  • other 92.2% (fraction of BRG1 peaks following linear response model)
  • other 44.2% (BRG1-bound enhancers with decreased accessibility)
  • other 16.1% / 17.8% (BRG1-bound promoters / insulators with decreased accessibility)
  • other 83.2% (DARs located at enhancers)
  • other 50.7% linear vs 49.3% buffered (distribution of BRG1-dependent regulatory element responses)
  • other 78% protein remaining at 0.3 nM; 12% at 10 nM dTAG13 (BRG1 protein levels at given dTAG13 doses)
  • other 23.9% / 65.0% / 11.1% (BRG1 binding distribution at promoters / enhancers / insulators)

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 uses a targeted protein degradation (dTAG) system to titrate BRG1 protein across five concentrations in mouse embryonic stem cells, profiling chromatin binding (CUT&RUN), accessibility (ATAC-seq), and transcription at each level. A custom delta-value framework was developed to classify individual genomic loci as exhibiting linear or buffered dose-response kinetics. Enrichment of TF motifs and genomic element types across response categories was assessed by two-sided Fisher's exact tests. Quantitative protein measurements from Western blots were summarized as mean ± SEM across six biological replicates.

Replicationbiological Sample sizeSix independent biological replicates for Western blot quantification; three independent experiments for SNF2H depletion Western blot; replicate counts for CUT&RUN and ATAC-seq experiments not stated in the provided text GroupsFive dTAG13 concentrations (DMSO/0 nM through 10 nM) representing graded BRG1 depletion; BRG1-dependent vs. BRG1-independent regulatory elements; Brg1-knock-in vs. wild-type mES cells; single vs. dual remodeler depletion (BRG1, SNF2H) Pairingunclear Randomization/blindingnot stated DispersionSEM Exact p-valuesno Effect sizesyes Confidence intervalsno Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Two-sided Fisher's exact test TF motif enrichment at BRG1-independent versus BRG1-dependent ATAC-seq peaks in promoters, enhancers, and insulators — Fig. 3a not stated
Two-sided Fisher's exact test Enrichment of genomic element types per delta-value response group versus all consensus background regions — Fig. 4d not stated
Custom delta-value classification (distance to idealized linear vs. buffered model) Classification of BRG1 CUT&RUN binding peaks and ATAC-seq accessibility peaks across five dTAG13 concentrations — Figs. 1i–k, 4a–b 20,326 BRG1-binding sites stated for Fig. 1k histogram not stated
Principal component analysis (PCA) Dimensionality reduction of BRG1 CUT&RUN signal matrix across dTAG13 concentrations — Fig. 1f na
Approaches that could also have been used
  • Dose-response kinetics at each genomic locus were classified as 'linear' or 'buffered' using a custom delta-value metric measuring distance to two idealized prototype curves
    Could also: Parametric dose-response curve fitting (e.g., four-parameter logistic or Hill equation applied per locus) could also be used — Curve fitting would yield continuous, interpretable parameters per locus — such as EC50 and Hill coefficient — enabling quantitative ranking of sensitivity rather than binary classification, and uncertainty in each parameter could be captured with confidence intervals
  • Multiple Fisher's exact tests were performed across many TF motifs without an explicitly stated multiple-testing correction
    Could also: Benjamini–Hochberg FDR correction applied across the full family of motif tests would also be a standard approach — When testing hundreds of motifs simultaneously, FDR control defines the expected proportion of false positives among declared significant calls, which helps readers gauge the biological specificity of enriched motifs
  • Protein abundance from Western blots was summarized as mean ± SEM across six biological replicates
    Could also: SD or a 95% confidence interval could also be used to express dispersion — SD describes the biological variability of individual measurements, while a 95% CI conveys precision of the mean estimate; both are straightforward to interpret for readers assessing reproducibility at n=6, and SD is generally preferred when characterizing raw biological spread rather than estimation error
  • TF motif enrichment was assessed independently per motif with Fisher's exact test, treating each motif as a separate binary comparison
    Could also: A multi-variable logistic regression or regularized approach (e.g., LASSO) applied jointly across motifs could also be used — A joint model estimates each motif's independent contribution while controlling for co-occurring motifs — relevant because TF binding sites are often correlated — and naturally yields a single significance threshold across the feature set
  • Global shifts in BRG1 binding across concentrations were visualized using PCA of the CUT&RUN signal matrix
    Could also: Hierarchical clustering of samples or UMAP could also be used to represent sample-level relationships in high-dimensional space — UMAP can reveal non-linear structure not captured by the first two principal components; hierarchical clustering provides a dendrogram that directly encodes pairwise similarity and is easy to interpret alongside heatmaps
  • The linear-versus-buffered classification used a deterministic geometric delta value with no statistical uncertainty estimate
    Could also: A permutation test or bootstrap procedure could also be applied to assign a confidence level to each locus's classification — Bootstrapping the delta values would quantify how stably each locus is assigned to its category given the variability in the underlying signal, which is particularly informative for loci near the delta = 0 boundary between linear and buffered classes
Software: not stated in provided text

Citation network

Where this publication sits in the reproducibility-weighted citation graph — what it is built on, and what is built on it. Citation data from OpenAlex.

Citations
4
Impact: low
Foundation confidence
None of its references are in our reproducibility record yet — its foundation cannot be assessed.
Topics

No assessed neighbours yet — the network grows as more papers are assessed.

Data lineage

The datasets this paper uses (text-mined from the full text via Europe PMC), and which other assessed papers stand on the same data. A shared dataset is a factual link — not a judgement.

GSE30206 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet

What was reproduced

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

scope.md — pmid-40846763

Paper: Hagihara, Zhang & Zhang. Precise modulation of BRG1 levels reveals features of mSWI/SNF dosage sensitivity. Nat Genet 57:2250–2263 (2025). PMID 40846763 · PMCID PMC12425804 · DOI 10.1038/s41588-025-02305-z.

Linked artifacts (from the brief):

  • Code: github.com/guifengwei/ChromHMM_mESC_mm10 — this is NOT the paper's own code. It is a third-party precomputed resource: a 12-state ChromHMM chromatin state map for mouse ESC (E14, mm10), built by Guifeng Wei (2015) from ENCODE ChIP-seq. The paper downloaded this map (Methods: "A chromatin states map for mES cells … was downloaded from https://github.com/guifengwei/ChromHMM_mESC_mm10").
  • Data: geo:GSE30206text-mining mismatch. GSE30206 = "DNA binding factors shape the mouse methylome…" (Stadler 2011), a methylome SuperSeries unrelated to this paper. The paper's own data are GSE274469 + GSE294015; its own code is github.com/YiZhang-lab/Brg1_dTag (Zenodo 10.5281/zenodo.15951452).

How the paper uses ChromHMM (the in-scope pipeline)

Methods: the downloaded 12-state map was used to compute chromatin-state enrichments of given genomic regions (the BRG1-dependent enhancer groups G1–G5) using Fisher's exact test vs. a background of bedtools-shuffled regions (shuffleBed). Reported in Fig. 4d/4e (and the SE-overlap statements in the text: "231 SEs … 97 (42%) classified as G1 … 195 (84.4%) among G1–G5").

In scope vs out of scope

Result Pipeline In scope? Why
Properties of the shipped 12-state map (n states, functional categories, per-state genome coverage, dense↔segments consistency) ChromHMM map (shipped BED) — deterministic YES (primary) Fully shipped in the repo; deterministic; directly checks the paper's description of the resource.
Paper's described enrichment method runs & behaves correctly (Fisher vs shuffleBed background) on the shipped map bedtools + Fisher YES (secondary, method face-validity) Reproduces the exact procedure the paper describes, using the shipped map + a public deterministic region set (mm10 TSS).
Fig. 4d/4e: chromatin-state enrichment of enhancer groups G1–G5 full ATAC→DAR→delta-clustering→enrichment NO (the hard 20%) (a) No source-data values are reported for Fig 4d (the Source-Data workbook covers Fig 4c/4g, not 4d) → nothing to pin a 1:1 number to. (b) The G1–G5 enhancer coordinates are not deposited — GEO ships only per-sample bigwigs (no peaks/DARs/beds); the authors' R code references intermediates on their local filesystem (/nfs4/chaozhang/...). Reproducing them requires rebuilding the entire upstream pipeline from bigwigs.
Re-training the ChromHMM map itself (LearnModel, 12 states) ChromHMM LearnModel on ENCODE marks NO Non-deterministic (random init → state numbering won't match); underspecified ENCODE input list/read-processing; this is the third-party resource, not a paper result.
Wet-lab (CRISPR, WB, IF, growth assays), ATAC/CUT&RUN peak-calling, bulk-RNA DEGs, ChromBPNet, BEAS-2B various NO Out of scope (wet-lab) or separate pipelines not tied to the linked ChromHMM resource.

Plan

Run on «our HPC»/«infra» (all data stays on «infra»; only small result tables return):

  1. git clone the guifengwei map repo on «infra»; gunzip mESC_E14_12_segments.bed and mESC_E14_12_dense.annotated.bed.
  2. Compute: number of distinct states; per-state genome coverage (bp, %) and segment counts; extract the dense state→functional-name map; categorise the 12 states into the paper's 6 functional categories; check dense↔segments consistency.
  3. Reproduce the paper's enrichment method: download mm10 chrom.sizes + RefSeq TSS (UCSC, on the compute node); intersect TSS with the state map; enrichment of TSS in promoter-labelled states vs. bedtools shuffle background, Fisher's exact test — confirming the pipeline is runnable and the map's
C1
Reported
A 12-state model was trained (Methods, ChromHMM)
Reproduced
12 states (E1-E12) in the shipped segmentation; 12 named states in the dense annotation
exact
C2
Reported
states cover promoters, enhancers, repressed Polycomb, transcription elongation, CTCF-bound, intergenic (6 categories)
Reproduced
all 6/6 categories present in the shipped dense state names
exact
C3
Reported
(no per-state coverage number stated in paper)
Reproduced
full per-state genome coverage table; dense total bp == segments total bp (diff 0), i.e. dense is a faithful re-encoding of the segmentation
partial
C5
Reported
enrichment method = Fisher's exact vs bedtools-shuffleBed background (Fig 4d/4e); active-promoter states expected strongly enriched
Reproduced
method runs on the shipped map; 7_ActivePromoter 63.7x enriched at mm10 TSS (Fisher OR=117, p~0, 21723/47248), BivalentChromatin 23.1x, StrongEnhancer 4.2x; intergenic/heterochromatin depleted
partial

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 75/100

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)
Scoring basis — itemised

Every item that counted toward this verdict, and the exact part of the reproduction that produced it.

Supporting (toward a concern)
Content-critical question only partially held
+2 pts
From: Q5 · Derivability / plausibility 🟡
Content-critical question only partially held
+2 pts
From: Q7 · Core claim 🟡
Content-critical question only partially held
+2 pts
From: Q8 · Severity of the miss (overall human judgment) 🟡
Minor / cosmetic deviation
+1 pts
From: Q4 · Cause of the deviation 🟡
Minor / cosmetic deviation
+1 pts
From: Q2 · Endpoint comparability 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score +6

The checkable descriptive claims reproduced exactly — 12 states (C1), all 6/6 functional categories (C2), and a dense annotation that faithfully re-encodes the segmentation (dense bp == segments bp, diff 0, C3) — and the paper's enrichment method behaves sensibly on the shipped map (7_ActivePromoter 63.7× at TSS, Fisher OR=117, p≈0, C5). The one shortfall is on data-availability/authors' side: the headline Fig 4d/4e chromatin-state enrichment of enhancer groups G1–G5 is not 1:1 reproducible because those coordinates are not deposited (GEO ships only bigwigs) and no source-data numbers are reported. There is no factual deviation in anything that could be compared and no fabrication concern — this is an auditability gap, not a discrepancy — so the case is solid but limited.

🤝
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.

207.2 k
tokens (I/O) · 15.2 M incl. cache
22 min
runtime · 0 CPU-h
0.3 GB
peak RAM
2
HPC jobs
hummel
machine