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

Gli1-expressing stromal cells are highly reparative precursors of long-lived chondroprogenitors in the fetal murine limb.

Nat Commun · 2025
L1 10/100 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 result did not reproduce in this reproduction attempt. Where our recomputation produced values that differ from the published ones, those discrepancies are listed below. This is a single automated attempt — not peer review and not a finding of error or misconduct — and differences can also arise from data access, undocumented parameters or the computing environment. The verdict can be contested via “report an error”.

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: Q3 · Location of the main deviation 🟡
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 +7
✓ What held up
  • Nothing in this column.
What did not (or only partly)
  • 🟡Could not use the authors’ exact input data
  • 🟡Reported values were only indirectly comparable
  • 🟡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
10/100
Reproducibility score
3.6 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 0% of all assessed papers rank 1169 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 to RUN the method, but the reported RESULT did not reproduce 1:1. The brief's code link is the generic third-party tool github.com/Oshlack/speckle (propeller cell-type-proportion test), valid under P16. Data GSE273540 (8 scRNA-seq 10x samples, 4 control vs 4 p21-injury E14.5 mouse limb) downloads cleanly; only RAW 10x matrices are deposited -- NO authors' annotated Seurat object. I therefore re-derived cell types independently (Seurat5 merge of 67,280 QC'd cells -> Harmony over sample -> 18 clusters at res0.5 -> marker-panel argmax annotation) and ran speckle::propeller(experimental vs control, logit transform). The propeller pipeline ran end-to-end on the paper's own data, so the METHOD is reproducible. But the paper's qualitative claim 'chondrocyte proportion increased in experimental' is MISMATCHED: in my annotation chondrocytes are slightly LOWER in experimental (0.386->0.322) and NOTHING is significant (min FDR=0.434, threshold 0.1). Honest interpretation, flagged for the human reviewer: (1) the authors ship no annotation, so the disagreement may be annotation-driven, not a real discrepancy; (2) the paper's specific increase is about the Gli1-LINEAGE-TRACED subset, which needs transgenic lineage labels not derivable from standard matrices -- declared OUT OF SCOPE; the whole-dataset propeller that Methods describe does not show the increase; (3) n=4/group with high control heterogeneity (chondrocyte 0.279-0.496) is underpowered. NOT attempted (the hard last 20% / out of scope): CellBender + DoubletFinder exact QC, lineage-traced Gli1/Pdgfra subset proportions, snRNA-seq, all wet-lab/IHC. No fabrication signal -- the reported claim is qualitative (no printed p/FDR) and concerns a subset not reconstructed here.

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 10
    assessed: 2026-06-16 ⛓ 3257169cc52a
✎ 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
2026-06-16
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-16
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 paper asks whether postnatal long-lived cartilage progenitors (LLCPs) arise from fetal short-lived progenitors or are already present (or recruited from outside the cartilage) and activated later, and whether such progenitor behaviour can adapt to developmental perturbations to confer growth robustness.

Core claims
  • Fetal Gli1+ cells (including cells outside the cartilage) are the precursors of postnatal long-lived chondroprogenitors and give rise to most growth-plate chondrocytes juvenile/adult. finding
  • Gli1+ LLCP precursors remain mostly dormant until postnatal stages but expand in the cartilage in response to genetically-induced cell-cycle arrest, enabling normal growth (compensation). finding
  • Gli1-derived chondrocytes are required to compensate for cell-cycle arrest in the fetal cartilage. mechanism
  • Reparative Gli1+ cells originate from Pdgfra+ cells outside the cartilage, revealing surrounding tissues as a chondroprogenitor source. finding
  • Gli1 expression can be activated de novo in cells that were previously Gli1-negative after p21 induction, revealing cellular plasticity. finding
  • Single-nuclei RNA-seq with SCENIC regulon analysis identifies upregulated Gli1 transcription factor activity in experimental osteochondroprogenitors. method
  • Combined Cre/rtTA inducible systems (Pan-Cart-p21MOE, Gli-Cart-DTA) allow cartilage-targeted p21 overexpression and selective ablation of Gli1-derived chondrocytes. method
Experimental setups
Assay System Perturbation Readout Platform
single-nuclei RNA-seq (snRNA-seq) fetal mouse left and right knees (Left-Cart-p21MOE and control embryos) cartilage-targeted mosaic p21 overexpression (Pitx2-Cre; Col2a1-rtTA; Tigre Dragon-p21) nuclear transcriptomes; cell clustering; regulon (Gli1) activity via SCENIC
bulk RNA-seq (prior dataset, reanalysed) Left-Cart-p21MOE mouse cartilage cartilage-targeted p21 overexpression gene/regulon expression
Cre-based lineage tracing (tdTomato reporter) mouse fetal/postnatal limb cartilage and surrounding tissue Gli1CreER lineage tracing ± Pan-Cart-p21MOE (Dox-induced p21), tamoxifen induction distribution/number of tdTom+ cells inside/outside cartilage at 1,4,7 days post-TM and P0
long-term lineage tracing mouse growth plate (E14.5 to P60) Gli1CreER lineage traced from E12.5 or E13.5 % tdTom+ chondrocytes over time (E14.5, E17.5, E19.5, P30, P60)
multicolour clonal lineage tracing (RGBow reporter) mouse growth plate cartilage Gli1CreER, TM at E12.5 or E13.5 number, location, and length of clonal columns across cartilage zones RGBow tri-colour reporter line
EdU/CldU label-retaining and proliferation assay Gli1CreER; R26 LSL-tdTom mouse RZ cartilage EdU at E13.5/E14.5 (60 mg/kg), CldU at collection; TM at E12.5 EdU-retaining cells and proliferative activity at P3, P7, P14
genetic cell ablation (TUNEL apoptosis) mouse fetal cartilage (Gli-Cart-DTA, Pan-Cart-p21MOE; Gli-Cart-DTA) inducible attenuated diphtheria toxin (Tigre Dragon-aDTA) to kill Gli1-derived chondrocytes ± p21 TUNEL apoptosis; P0 bone (mineralised region) length
bone length/morphometry postnatal mouse tibia and femur (to P100) Left-Cart-p21MOE vs control left/right bone length ratio
Key results
  • Gli1 transcription factor regulon activity significantly upregulated in left experimental osteochondroprogenitors and to lesser extent resting chondrocytes.
  • Gli1-lineage showed significant cartilage expansion in Pan-Cart-p21MOE vs controls at all stages analysed (1,4,7 days, P0).
  • Gli1-lineage cells contribution to cartilage increased over time, labelling up to 65-70% of chondrocytes in the 2-month-old growth plate. 65-70%
  • Only Pan-Cart-p21MOE; Gli-Cart-DTA pups showed significantly decreased bone length vs Pan-Cart-p21MOE and/or controls, showing Gli1-derived chondrocytes required for compensation.
  • Left/right bone length ratio not significantly different for tibia and only ~1.5% lower for femur in Left-Cart-p21MOE to P100; no major asymmetries. ~1.5% lower (femur)
  • Gli1 mRNA detected in less than 40% of E13.5-traced chondrocytes at E14.5 and E17.5; more lineage-traced chondrocytes retained Gli1 in Pan-Cart-p21MOE. <40%
  • Cells initially Gli1-negative that gained Gli1 expression increased in experimental condition; significant at E17.5 (5 days post-Dox) but not E14.5.
  • EdU-retaining cells in RZ became enriched in Gli1-lineage between P7 and P14, with a proliferation peak at P7.
Key statistics
  • count ~32,000 nuclei in 32 clusters (original snRNA-seq dataset)
  • count ~21,000 nuclei, 23 clusters (lateral-plate-mesoderm-derived subset)
  • percent 65-70% of chondrocytes (Gli1-lineage labelling in 2-month-old growth plate)
  • percent 60-80% of chondrocytes (prior study: P30 Gli1+ cells label growth-plate chondrocytes 1 month later)
  • percent 60-70% of E14.5 chondrocytes express p21 (Pan-Cart-p21MOE induction validation)
  • percent <40% (E13.5-traced tdTom+ chondrocytes detectably expressing Gli1 mRNA)
  • percent ~1.5% lower (femur left/right length ratio, Left-Cart-p21MOE vs control)
  • count n = 17 Ctl, 9 Exp; n = 17, 9, 6, 10 (bone length/growth defect cohorts)

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 combines mouse genetic lineage tracing (Cre-based with tamoxifen induction), single-nucleus RNA sequencing (snRNA-seq with SCENIC regulon analysis on ~21,000 nuclei from 4 embryos), and quantitative histology to characterise Gli1+ progenitor populations during fetal and postnatal skeletal growth. Group comparisons of cell proportions, lineage contributions, and bone lengths were made primarily using one-way or two-way ANOVA with Tukey HSD or Sidak post-hoc corrections. Results are reported as mean ± SD, with p-values displayed on graphs selectively when ≤0.1 and full ANOVA results tabulated in supplementary material.

Replicationbiological Sample sizeSample sizes stated per figure panel per genotype group; no formal power analysis mentioned in this excerpt GroupsControl vs experimental genotypes (p21 overexpression variants, DTA ablation, lineage-tracing combinations); multiple developmental time points (E14.5 through P60) Pairingmixed Randomization/blindingnot stated DispersionSD Effect sizesno Confidence intervalsno Multiplicity correctionTukey HSD and Sidak's correction (each used in different figure comparisons)
Statistical tests used
Test Applied to n Assumptions
One-way ANOVA with Tukey HSD post-hoc correction Gli1 regulon activity across cell types and conditions from SCENIC analysis (Fig. 1a'; Supplementary Table 1) ~21,000 nuclei (lateral-plate-mesoderm-derived subset) from 2 control and 2 experimental embryos not stated
Two-way ANOVA P0 bone length comparison between control and Pan-Cart-p21 MOE genotypes (Fig. 1b') n=17 Ctl, 9 Exp not stated
Two-way ANOVA with Sidak's multiple comparisons post-hoc Gli1-lineage proportion inside/outside cartilage across three time points (Fig. 1c') n=11, 6, 11 Ctl; 9, 6, 4 Exp across three time points not stated
ANOVA with Tukey's post-hoc TUNEL-positive cell quantification across four genotypes at E17.5 and P0 (Fig. 2b') n=8, 6, 6, 6 at E17.5; n=6, 5, 7, 10 at P0 not stated
Two-way ANOVA with Sidak's multiple comparisons post-hoc P0 mineralised region length across four genotypes (Fig. 2c) n=17, 9, 6, 10 not stated
Bioinformatics cell-proportion analysis (method cited by reference number, not named in text) Proportion of Gli1-high cells assigned to chondrocyte vs other clusters, experimental vs control limbs (Supplementary Fig. 2e) ~21,000 nuclei from 2 control and 2 experimental embryos not stated
Approaches that could also have been used
  • Two-way ANOVA with Sidak's or Tukey's post-hoc was used for multi-group comparisons of bone length and cell proportions, treating embryos/pups as independent observations
    Could also: A linear mixed model (LMM) with litter as a random effect could also be used — Embryos from the same litter share a maternal environment and genetic background; an LMM accounts for this within-litter correlation, which when ignored can underestimate standard errors and inflate type-I error rates
  • snRNA-seq cell-type proportion comparisons between conditions were based on n=2 embryos per group
    Could also: Dedicated compositional methods such as scCODA or Dirichlet regression, applied to n≥3 biological replicates per group, could also be used — Cell-type proportions are compositional (sum to 1); dedicated frameworks explicitly model this constraint and estimate between-sample variance, which requires at least three replicates per group for stable inference
  • Dispersion is reported as mean ± SD throughout all quantitative figure panels
    Could also: 95% confidence intervals could also be plotted alongside or instead of SD — For group sizes of n=4–11, 95% CIs directly convey uncertainty about the group mean estimate, which is complementary to SD's description of data spread and is increasingly requested by journals for small-n comparisons
  • p-values are shown on graphs only when ≤0.1, leaving non-significant comparisons unlabelled
    Could also: Reporting all exact post-hoc p-values regardless of threshold is also standard practice — Showing all values lets readers assess the complete pattern of evidence and avoids ambiguity about whether unlabelled comparisons were tested and non-significant or simply not shown
  • Effect sizes are not reported alongside ANOVA p-values for any of the quantitative comparisons
    Could also: Standardised effect sizes (e.g., partial η², Cohen's d) could also be reported — Effect sizes quantify the magnitude of differences independently of sample size, supporting biological interpretation and enabling future power calculations or meta-analyses
  • The left limb (experimental) vs right limb (control) within the same animal was used for the initial bone-length asymmetry analysis, with the ratio reported but a paired test not explicitly named
    Could also: A paired t-test or Wilcoxon signed-rank test on the left/right ratio per animal could also be applied explicitly — Within-animal pairing on the left/right axis directly removes between-animal variability; formally specifying the paired test makes the analytical choice transparent and may increase sensitivity for small asymmetries
Software: SCENIC (regulon activity analysis)

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
0
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.

RRID:Addgene_61580 RRID in Article (http://semanticscience.org/resource/SIO_001029)
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 — pmid-41253754

Paper: Qu et al. 2025, Nat Commun 16, "Gli1-expressing stromal cells are highly reparative precursors of long-lived chondroprogenitors in the fetal murine limb." DOI 10.1038/s41467-025-65029-y · PMCID PMC12627582.

Code link in brief: https://github.com/Oshlack/speckle — this is the generic third-party tool (speckle R package, propeller cell-type-proportion test), NOT an authors' analysis repo. Per brief rule P16, applying this tool to the paper's data is a valid reproduction.

Data: GEO GSE273540 — 8 scRNA-seq samples (10x), Mus musculus E14.5 limb: 4 control (L/R limbs, 2 control embryos) + 4 experimental (L/R limbs, p21-MOE injury, embryos 2&3). GEO ships raw 10x matrices only (GSE273540_RAW.tar, MTX

  • TSV) — no annotated Seurat object / cell-type labels are deposited.

In scope (pipeline-derived)

  • R1 — speckle/propeller cell-type proportion test, experimental vs control. Methods (verbatim): "To analyse the distinct cellular compositions among conditions, the propeller test available in the speckle R (v.0.0.1 …) was employed. The groups were classified with a false discovery rate (FDR) of ≤0.1 as indicative of notable variations in cell types." Reported result (Results, qualitative): "the proportion of cells belonging to the chondrocyte group was increased in experimental limbs, as compared to Ctl." → Reproduce: cell-type proportions per sample, run propeller, test whether the chondrocyte proportion is significantly increased in experimental (FDR≤0.1) and in the reported direction.

Required upstream (not the headline result, but needed to produce cell types)

  • Standard Seurat pipeline on the raw matrices: load 8 samples → QC → normalize → integrate (Harmony) → cluster → annotate broad cell types by canonical limb-mesenchyme markers (chondrocyte: Col2a1/Acan/Sox9/Col9a1; perichondrium; tenocyte: Scx/Tnmd; fibroblast: Pdgfra/Col1a1; joint/mesenchyme; plus muscle/endothelial/immune/blood). The authors' exact annotation is not shipped, so this is an independent re-derivation — the propeller grade will therefore be partial (direction + significance, not a bit-identical value).

Out of scope (not attempted, why)

  • Gli1/Pdgfra lineage-traced subset proportions ("within the Gli1high LPM- derived population…") — requires transgenic reporter/lineage labels not derivable from standard transcriptomes. Wet-lab.
  • snRNA-seq, IHC/imaging, in-situ, all wet-lab phenotyping.
  • CellBender ambient-RNA removal + DoubletFinder (paper's exact QC) — skipped as the optional last-20%; standard Seurat QC used instead. Noted as a deviation.

Comparison target

No numeric propeller value is reported in the main text (only direction + FDR≤0.1 threshold). Comparison is therefore: (a) chondrocyte proportion direction (UP in experimental) and (b) whether it reaches FDR≤0.1 under propeller. This is a qualitative/partial 1:1.

R1
Reported
chondrocyte group proportion increased in experimental (p21-MOE injury) vs control limbs (speckle/propeller, FDR<=0.1)
Reproduced
Chondrocyte PropMean Control=0.386 vs Experimental=0.322 (DECREASED in exp); propeller T=+1.55, P=0.146, FDR=0.434; no cell type reached FDR<=0.1 (min FDR=0.434 across 8 broad cell types)
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 10/100

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: Q3 · Location of the main deviation 🟡
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 +7

The speckle/propeller method is reproducible end-to-end on the paper's own GSE273540 data, but the reported result does not reproduce: whole-dataset chondrocyte proportion is lower in experimental (0.386->0.322) and nothing reaches FDR<=0.1 (min FDR=0.434), reversing both the direction and the implied significance. The deviation is factually severe (q6 red) yet most plausibly sits on the input/scope side — the authors deposited no annotated Seurat object so cell types were independently re-derived, and the paper's actual claim is about the Gli1-lineage-traced subset, which needs transgenic labels not present in standard matrices (out of scope). This is a data-availability + annotation issue, not a fabrication signal; with n=4/group and high control heterogeneity (0.279-0.496) the comparison is also underpowered, so overall it reads as a solid run with explainable, human-review-worthy deviations.

🤝
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.

163.5 k
tokens (I/O) · 12.4 M incl. cache
32 min
runtime
Per-job HPC accounting not captured for this run — the runtime shown is the reproduction’s measured wall-clock time.