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

Epigenome screening highlights that JMJD6 confers an epigenetic vulnerability and mediates sunitinib sensitivity in renal cell carcinoma.

Clin Transl Med · 2021
L1 85/100 3/4
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: Q8 · Severity of the miss (overall human judgment) 🟡
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 +1
✓ What held up
  • Same input data as the authors
  • No relevant deviation in data/preprocessing
  • No authors-side cause for any deviation
  • Reported values are derivable from the shared data
  • Any deviation was negligible
  • The central claim held under reproduction
What did not (or only partly)
  • 🟡Reported values were only indirectly comparable
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
85/100
Reproducibility score
0.6 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 67% of all assessed papers rank 348 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 PUBLIC part. Reproduced the paper's central epigenome-screening claim (JMJD6 over-expressed in RCC tumour vs normal, Fig 2A-D) on its own public GEO data using third-party tools (GEOquery+limma/base-R t-test, per rule P16). GSE40435 (the registry-assigned accession, 101 tumour/normal pairs): JMJD6 UP, p=6.6e-08 — direction+significance match, though effect size is modest (1.06x). GSE53757 (validation, 72 pairs): UP, p=4.5e-19, 1.65x. Grade within-tol because the paper states only 'significantly higher' (no exact number for these cohorts), so the comparison is directional+significance, not exact-value. NOT attempted (and why): the registry 'code' link BradnerLab/pipeline (ROSE2) and its reported super-enhancer counts (545 vs 286, Fig 6B), 2311 DEGs, 56531 ChIP peaks, and 1904-gene signature (Fig 5) all depend on the authors' OWN H3K27ac ChIP-seq / RNA-seq, which is 'available from the corresponding author upon reasonable request' and not deposited -> data_restricted, out of scope. TCGA-KIRC survival HR=1.289 (Fig 2J) is public but the LASSO/multivariate-Cox covariate specification is under-specified -> deferred as the optional hard 20%. IHC/organoid/CRISPRi/sunitinib results are wet-lab (non_pipeline). Overall: a clean partial 1:1 on the public screening claim; the headline mechanistic super-enhancer results are unverifiable without the restricted raw data.

💻 Code ↗ 🗄 Data: GSE40435

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 85
    assessed: 2026-06-15 ⛓ 34c683ba190e
✎ 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
2026-06-15
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

The study tests whether integrating CRISPR/Cas9 functional screening with RCC genomic datasets can reveal a novel epigenetic vulnerability, hypothesizing that JMJD6 is a pivotal chromatin modifier that drives renal cell carcinoma progression and represents a therapeutic target whose inhibition sensitizes RCC to sunitinib.

Core claims
  • JMJD6 is an epigenetic vulnerability/fitness gene in RCC, identified by integrating GeCK CRISPR screening data with TCGA/ICGC RCC cohorts. finding
  • High JMJD6 expression correlates with poor survival and promotes RCC progression in vitro and in vivo. finding
  • Aberrant p300 drives high JMJD6 expression, which assembles super-enhancers to activate identity/oncogenic genes including VEGFA, β-catenin, and SRC. mechanism
  • The JMJD6 inhibitor SKLB325 suppresses JMJD6-mediated oncogenic effects in RCC cells, patient-derived organoids, and in vivo. finding
  • Targeting JMJD6 sensitizes RCC to sunitinib and is synergistic in combination. finding
  • Integration of GeCK pan-cancer fitness genes with univariate Cox-screened epigenetic regulators yielded 61 candidate epigenetic fitness genes, prioritized by AUC and validated by MTT/siRNA. method
  • ChIP-seq with ROSE2 was used to identify JMJD6-assembled super-enhancer regions from distal H3K27ac peaks. method
  • JMJD6 is highly expressed in kidney cancer relative to most other solid tumors (CCLE and 33-tumor pan-cancer analysis). finding
Experimental setups
Assay System Perturbation Readout Platform
CRISPR/Cas9 functional knockout screen (GeCK) data integration 33 tumor cell lines including RCC; TCGA-KIRC and ICGC-RCC cohorts genome-wide knockout fitness/essential genes for tumor growth
MTT / CCK-8 cell proliferation assay RCC cell lines (786-O, ACHN, Caki-1) siRNA knockdown of candidate epigenetic genes / JMJD6 cell proliferation/viability (absorbance at 450 nm) Bio-Rad microplate reader; Dojindo CCK-8
siRNA knockdown / Western blot 786-O and ACHN RCC cells 20 nM siRNA knockdown of JMJD6 knockdown efficiency / protein levels ECL chemiluminescence (Santa Cruz)
CRISPR single-cell JMJD6 knockout clone generation 786-O cells JMJD6 KO via pX459 knockout efficiency (Western blot, Sanger sequencing) pX459 plasmid
Colony formation and Transwell migration/invasion assays RCC cell lines (786-O, ACHN) JMJD6 knockdown/knockout colony number; migration/invasion cell counts Costar transwell; BD Matrigel/fibronectin
RNA-seq (transcriptome) RCC cells JMJD6 modulation differentially expressed genes (DEGs) HiSeq (BGI); aligned hg19/Bowtie, RSEM
ChIP-seq and ChIP-qPCR RCC cells none / p300 IP H3K27ac super-enhancer regions (ROSE2); p300 enrichment at JMJD6 promoter Active Motif EpiShear sonicator; Active Motif Inc.
Xenograft / PDX / patient-derived organoid models BALB/c nu/nu mice; human ccRCC tissues JMJD6 modulation; SKLB325; sunitinib tumor growth; organoid number/diameter BD Matrigel
Key results
  • Overlap of GeCK fitness genes and hazardous epigenetic regulators identified 61 candidate epigenetic fitness genes in RCC. 61 genes
  • MTT assay identified JMJD6 as the most potent hit among candidate epigenetic factors in RCC cells.
  • JMJD6 knockdown suppressed RCC proliferation across three independent RCC cell lines.
  • JMJD6 is highly expressed in kidney cancer relative to most other solid tumors and in RCC vs other tumor types.
  • JMJD6 assembles super-enhancers driving kidney cancer identity genes including VEGFA, β-catenin, and SRC.
  • SKLB325 suppressed JMJD6-mediated oncogenic effects in cells, organoids, and in vivo.
  • Targeting JMJD6 sensitized RCC to sunitinib with synergistic effect in combination.
Key statistics
  • count 1614 fitness genes in RCC (derived from pan-cancer GeCK screening results)
  • count 665 epigenetic regulators (expression data from TCGA-KIRC cohort)
  • count 355 hazardous epigenetic factors (univariate Cox regression at P < 0.05)
  • count 61 potential epigenetic fitness genes (overlap of two screening results)
  • pvalue P < 0.05 (univariate Cox regression significance threshold for screening epigenetic factors)
  • count 33 tumors (pan-cancer JMJD6 expression analysis)
  • count 73,750 new cases; 14,830 deaths (2020 estimated US kidney cancer statistics)
  • other >30% (RCC cases progressed to terminal stage at diagnosis)

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 paper integrated CRISPR/Cas9 functional screening data with TCGA-KIRC and ICGC-RCC patient cohort transcriptomics to nominate JMJD6 as an epigenetic target in renal cell carcinoma. Differential expression in cohort data was assessed with Wilcoxon tests, survival associations via univariate Cox regression and Kaplan–Meier analysis, and gene selection via LASSO regression; RNA-seq DEGs were called using a Poisson distribution method. In vitro and in vivo experimental comparisons used one-way ANOVA or two-tailed Student's t-tests, with results expressed as mean ± SD across three biological replicates and significance defined at P < 0.05.

Replicationbiological Sample sizeThree biological replicates stated for all cell-line experiments; 5 mice per group for in vivo xenograft; cohort sizes from TCGA-KIRC and ICGC-RCC not specified in the provided text GroupsJMJD6 knockdown/knockout vs control; SKLB325 vs vehicle; SKLB325 + sunitinib combination vs single agents; RCC tumor vs matched normal tissue; JMJD6-high vs JMJD6-low in patient cohorts Pairingmixed Randomization/blindingnot stated DispersionSD Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionno
Statistical tests used
Test Applied to n Assumptions
Two-tailed Student's t-test Pairwise comparisons between treatment regimens across cell-line and in vivo experiments (proliferation, migration, invasion, drug-response assays) not stated
One-way ANOVA Multi-group comparisons between treatment regimens in cell-line and in vivo experiments not stated
Wilcoxon test Differential analysis of mRNA/protein levels in TCGA-KIRC and ICGC-RCC cohorts; pan-cancer differential analysis across 33 tumor types na
Univariate Cox regression Screening 665 epigenetic regulators for association with overall survival in TCGA-KIRC; 355 hazardous factors retained at P < 0.05 not stated
LASSO regression (glmnet) Prognostic gene selection from the candidate epigenetic regulators not stated
Kaplan–Meier analysis Overall survival analysis of RCC patients stratified by JMJD6 expression level na
AUC analysis Predictive efficiency evaluation of candidate prognostic genes in TCGA-KIRC and ICGC-RCC cohorts (Figure 1C) not stated
Poisson distribution method Detection of differentially expressed genes from RNA-seq count data (via RSEM/BGI pipeline) 3 biological replicates pooled per condition not stated
Approaches that could also have been used
  • RNA-seq differentially expressed genes were detected using the Poisson distribution method
    Could also: Negative binomial models as implemented in DESeq2 or edgeR's exactTest/glmLRT could also be applied to RNA-seq count data — Negative binomial models explicitly account for overdispersion (variance exceeding the mean) commonly observed in RNA-seq count data; they are a widely adopted standard and may yield better-calibrated p-values in settings where the Poisson equality of mean and variance does not hold
  • Univariate Cox regression was applied to each of 665 epigenetic regulators individually, retaining those with P < 0.05 without multiplicity correction
    Could also: Benjamini-Hochberg FDR correction across all 665 tests, or a LASSO-penalized Cox regression as a single joint model for the initial screening step, could also be used — When many hypotheses are tested simultaneously, FDR control reduces the expected proportion of false positives in the selected set; LASSO-Cox additionally handles correlated predictors jointly rather than one at a time, which can be advantageous when many epigenetic regulators co-vary
  • Multiple pairwise group comparisons were performed using separate Student's t-tests across many experiments
    Could also: One-way ANOVA followed by a post-hoc test with family-wise error rate control (e.g., Tukey HSD or Dunnett's test vs a shared control) could also be used when three or more groups are compared within a single experiment — An omnibus ANOVA with a post-hoc correction explicitly controls the family-wise error rate across all pairwise comparisons within an experiment, complementing the per-comparison t-test approach
  • Results were summarized as mean ± SD with significance indicated only by threshold symbols (*, **, ***, ****)
    Could also: Reporting exact p-values and a standardized effect size (e.g., Cohen's d or percent change with 95% CI) would also be standard practice — Exact p-values enable downstream meta-analysis and replication power calculations; effect sizes with confidence intervals convey the magnitude and precision of differences beyond a binary significance threshold and are increasingly expected by journals and systematic reviewers
  • Kaplan–Meier curves stratified by JMJD6 expression level were the primary survival presentation
    Could also: A multivariate Cox proportional hazards model adjusting for established clinical covariates (e.g., tumor stage, grade, age) could also accompany the Kaplan–Meier analysis — Multivariate Cox regression estimates the independent prognostic contribution of JMJD6 after accounting for known confounders, which is standard in clinical prognostic studies to assess whether the association is attributable to JMJD6 expression versus correlated clinical variables
  • Longitudinal CCK-8 proliferation measurements collected from Day 2 to Day 8 were analyzed as discrete time points
    Could also: A repeated-measures ANOVA or linear mixed-effects model treating time as a within-unit factor could also be used to model the full growth curve jointly — Measurements on the same experimental unit across time are correlated; a mixed-effects or repeated-measures model captures this structure and estimates a time-by-treatment interaction in a single test rather than requiring separate comparisons at each day
Software: R / RStudio 3.6.1 · edgeR · glmnet · survival · corrplot · ROSE2 (Rank Order of Super Enhancers) · Bowtie · RSEM

Result convergence & founder nodes

Findings this paper shares with others that ran a comparable experiment. A node’s strength is how many independent papers report it (replication breadth) — not how often it is cited, so a heavily-replicated but under-cited founder still stands out.

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
26
Impact: medium
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.

What was reproduced

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

Scope — pmid-33634984

Paper: Zhang et al. 2021, Clin Transl Med 11(2):e328. "Epigenome screening highlights that JMJD6 confers an epigenetic vulnerability and mediates sunitinib sensitivity in renal cell carcinoma." PMID 33634984 / PMC7882098 / doi:10.1002/ctm2.328.

Code link in registry: https://github.com/BradnerLab/pipeline (= ROSE2, the Rank-Ordering-of-Super-Enhancers tool — a third-party tool the authors applied, not authors' own repo). Per brief rule P16 this is equally valid, but its inputs (H3K27ac ChIP-seq) are not public (see below).

Data link in registry: GEO GSE40435 — confirmed via NCBI: Integrative genome-wide gene expression profiling of clear-cell RCC in Czech Republic, Illumina HumanHT-12 v4 (GPL10558), 101 ccRCC tumour / adjacent-normal pairs (202 samples). Public.

In scope (pipeline-derived, public data, specific reported claim)

# Reported result Paper loc Pipeline Data Feasibility
C1 JMJD6 expression significantly higher in tumour vs normal Fig 2A–D, Results 3.2 array DE / t-test (we apply limma — third-party tool on paper's data, P16) GSE40435 (assigned) HIGH — primary target
C2 same claim, validation cohort Fig 2A–D same GSE53757 (GPL570, 72 pairs) — also public HIGH — secondary

These reproduce the paper's central screening observation that JMJD6 is over-expressed in RCC tumour vs normal, using the exact public dataset named in the registry. The paper states the difference is "significant" (qualitative); we regenerate the direction, fold-change and exact p-value.

Out of scope — not attempted (with reason)

Result Paper loc Why dropped
545 vs 286 super-enhancers (ROSE2) Fig 6B Authors' own H3K27ac ChIP-seq of control vs JMJD6−/− cells — not deposited, "available from corresponding author upon reasonable request" → data_restricted
2311 DEGs (682 up / 1629 down) Fig 5A Authors' own RNA-seq of JMJD6−/− cells — not public → data_restricted
56,531 ChIP-seq peaks; 1904-gene JMJD6 signature Results 3.5, Fig 5G depend on the same restricted raw data
TCGA-KIRC survival HR=1.289 (1.136–1.463) Fig 2J Public (TCGA) but multivariate-Cox covariate set + LASSO feature selection under-specified; heavier; deferred as the optional hard 20%
Wet-lab: IHC 11/14 tissues, organoids, CRISPRi −60% VEGFA, sunitinib assays Fig 2–6 Manual/experimental, non_pipeline

Plan

Single «our HPC» SLURM job (std): build conda env (r-base + GEOquery + limma), download GSE40435 (+GSE53757) series matrices into «infra» inside the compute job, map JMJD6 probe(s), compute tumour-vs-normal t-test → small JSON pulled back to «host». No raw data on «host».

Figures / tables: Fig 2A
C1
Reported
JMJD6 significantly higher in RCC tumour vs normal in GSE40435 (Fig 2A-D, qualitative)
Reproduced
UP in tumor; FC=1.064 (log2FC=0.0899); Welch t=5.645; p=6.561e-08; n=101 tumour/101 normal
within tolerance
C2
Reported
JMJD6 significantly higher in RCC tumour vs normal in GSE53757 (Fig 2A-D, qualitative)
Reproduced
UP in tumor; FC=1.651 (log2FC=0.7231); Welch t=10.546; p=4.465e-19; n=72 tumour/72 normal
within tolerance

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 85/100

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: Q8 · Severity of the miss (overall human judgment) 🟡
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 +1

The public, in-scope screening claim is reproduced cleanly 1:1: JMJD6 is over-expressed in RCC tumour vs normal in both the registry-assigned cohort (GSE40435, p=6.56e-08) and the validation cohort (GSE53757, p=4.47e-19), matching the paper's qualitative 'significantly higher' statement in direction and significance, with no fabrication signal. The comparison is directional/significance-only because the paper prints no exact number for these cohorts (Fig 2A-D). However, the paper's headline mechanistic results (super-enhancer counts, DEGs, ChIP peaks, gene signature in Fig 5-6) rest on the authors' undeposited ChIP-seq/RNA-seq ('available upon request') and are therefore unverifiable — a data-availability limitation, not an authors' defect or a detected error. Overall: a solid partial reproduction.

🤝
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.

74.4 k
tokens (I/O) · 4.8 M incl. cache
10 min
runtime · 0.01 CPU-h
1.3 GB
peak RAM
1
HPC jobs
hummel
machine