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

Transcriptomic profiling of skeletal muscle adaptations to exercise and inactivity.

Nat Commun · 2020
L1 52/100 3/4
Why this verdict

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

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: Q6 · Severity 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 +8
✓ 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
52/100
Reproducibility score
1.3 SD below mean
vs. all fields · 1173 studies
🎯 Scores higher than 13% of all assessed papers rank 1014 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 reproduce the PIPELINE exactly, but the result is a 1:1 vs a LATER database version. MetaMEx's analysis code (annexes/functions.R::MetaAnalysis, a metafor::rma REML random-effects meta-analysis on shipped per-study limma/DESeq2 stats) was ported line-for-line and run on the repo's shipped per-study statistics (commit 651bbf1). CRITICAL CAVEAT: the GitHub repo ships database v3.2208 (2022); the paper reports v1 (2020), which is NOT deposited, so exact paper numbers are not expected. Result = PARTIAL: gene-level PPARGC1A reproduces near-exactly in two conditions (acute aerobic 2.24 vs 2.3-fold; inactivity -24.6% vs -25%); NR4A3 reproduces in DIRECTION in all conditions (headline 'most-regulated' gene), magnitudes drift with the larger v3.2 study set; genome-wide significant-gene counts are the same order of magnitude with acute-resistance (2354 vs 2404) and inactivity (1460 vs 1576) BH counts within ~2-7% of the paper. NOT ATTEMPTED (out of scope): the per-study limma/DESeq2 stage from raw GEO CEL/FASTQ (raw data + scripts not in repo; brief accession GSE4247 is invalid), and wet-lab/manual results. Reported paper values were auto-extracted from PMC full text and must be human-verified against the figures.

💻 Code ↗ 🗄 Data: GSE4247

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 52
    assessed: 2026-06-16 ⛓ beed7d1c2320
✎ 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 2026-06-18
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

Can a meta-analysis integrating publicly available skeletal muscle transcriptomic datasets identify the molecular pathways and genes selectively regulated by inactivity versus aerobic versus resistance and acute versus chronic exercise, and characterize the role of NR4A3 in exercise-induced metabolic responses?

Core claims
  • A meta-analysis (MetaMEx) of 66 published human skeletal muscle datasets reveals pathways selectively activated by inactivity, aerobic versus resistance, and acute versus chronic exercise training. finding
  • NR4A3 is one of the most exercise- and inactivity-responsive genes and mediates metabolic responses to exercise-like stimuli in vitro. mechanism
  • MetaMEx provides the most extensive resource of skeletal muscle transcriptional responses to exercise/inactivity with an online interface (www.metamex.eu) for interrogating the database. resource
  • Meta-analysis of transcriptomic data using restricted maximum likelihood integrates fold-changes across heterogeneous studies, yielding far more significant genes than individual studies. method
  • Acute exercise, exercise training, and inactivity segregate as distinct transcriptomic responses, with little difference between resistance and aerobic exercise. finding
  • MetaMEx highlights differential responses to exercise in individuals with metabolic impairments versus healthy individuals. finding
Experimental setups
Assay System Perturbation Readout Platform
transcriptomic meta-analysis (microarray/RNA-seq integration) human skeletal muscle (vastus lateralis, biceps brachii, quadriceps femoris) acute/chronic aerobic and resistance exercise and inactivity gene-level fold-change and FDR across studies
qPCR validation human skeletal muscle from independent healthy cohorts aerobic and resistance exercise mRNA expression of top exercise-responsive genes
in vitro metabolic characterization skeletal muscle cells (in vitro) exercise-like stimuli; NR4A3 manipulation metabolic responses
principal component analysis / correlation matrix / chord plot healthy human skeletal muscle datasets exercise modality and inactivity clustering and correlation of fold-changes across studies
Key results
  • PPARGC1A increased after acute aerobic exercise 2.3-fold (95% CI [1.6, 3.5])
  • PPARGC1A increased after acute resistance exercise 1.8-fold (95% CI [1.6, 2.2])
  • PPARGC1A decreased by inactivity 25%
  • PPARGC1A change greatest in biopsies taken after recovery period (>2h) versus immediately post-exercise 4.4-fold (95% CI [3.0, 6.4])
  • NR4A3 is the top exercise-responsive gene, strongly induced by acute aerobic and acute resistance exercise logFC 2.99 (aerobic); logFC 2.95 (resistance)
  • Acute aerobic and acute resistance exercise shared 360 commonly changed genes, while aerobic and resistance training shared only 25 360 vs 25 genes
  • Number of significantly responsive genes (FDR < 0.1%) per perturbation 897 acute aerobic; 2404 acute resistance; 1576 inactivity; 82 aerobic training; 2049 resistance training
  • qPCR validation in independent cohorts correlated highly with MetaMEx gene responses
Key statistics
  • fold_change 2.3-fold (95% CI [1.6, 3.5]) (PPARGC1A after acute aerobic exercise)
  • fold_change 1.8-fold (95% CI [1.6, 2.2]) (PPARGC1A after acute resistance exercise)
  • fold_change 4.4-fold (95% CI [3.0, 6.4]) (PPARGC1A in biopsies after >2h recovery)
  • fold_change logFC 2.99, FDR 2.0E-07 (NR4A3 acute aerobic)
  • fold_change logFC 2.95, FDR 8.2E-15 (NR4A3 acute resistance)
  • count 66 datasets / studies (total transcriptomic studies in MetaMEx)
  • count more than 1100 individuals (total subjects across meta-analysis)
  • count 897, 2404, 1576, 82, 2049 (responsive genes FDR<0.1% for acute aerobic, acute resistance, inactivity, aerobic training, resistance training)

Statistical methods review

Model: opus

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

This is a meta-analysis integrating 66 published human skeletal muscle transcriptomic datasets (>1100 individuals) across acute aerobic, acute resistance, aerobic training, resistance training and inactivity conditions. Per-gene effects were combined using a restricted maximum likelihood (REML) random-effects meta-analysis to compute pooled fold-changes and significance, with results reported as log fold-change, FDR and 95% confidence intervals (e.g., forest plots, M-plots, Venn diagrams). Exploratory inter-study structure was assessed with PCA, chord plots and fold-change correlation matrices, and top hits were further examined by qPCR in independent validation cohorts and by in vitro functional work on NR4A3.

Replicationbiological Sample sizeReported as number of studies per condition and total participants (>1100; per-condition male/female counts, age and BMI in Table 1); no formal power/sample-size calculation described GroupsAcute aerobic, acute resistance, aerobic training, resistance training and inactivity; healthy vs metabolically impaired individuals Pairingunclear Randomization/blindingna DispersionCI Exact p-valuesyes Effect sizesyes Confidence intervalsyes Multiplicity correctionFalse discovery rate (FDR) reported per gene; specific procedure not explicitly named
Statistical tests used
Test Applied to n Assumptions
Restricted maximum likelihood (REML) random-effects meta-analysis of per-gene effect sizes Pooled fold-change and significance for each exercise/inactivity-responsive gene across studies (Fig. 1, Fig. 2, Fig. 3, Table 3) 66 datasets / >1100 individuals; study counts per condition given in Tables 1-2 (e.g., 12 acute aerobic, 8 acute resistance, 7 inactivity, 11 aerobic training, 13 resistance training) not stated
Principal component analysis (PCA) Clustering of gene responses by intervention across healthy datasets (Fig. 2a) na
Correlation analysis of fold-changes (correlation matrix / chord plot) Inter-study similarity of fold-changes across all common genes (Fig. 2b, c) not stated
Correlation between qPCR validation-cohort responses and MetaMEx estimates Validation of top modality-specific genes in independent cohorts (Supplementary Fig. 2) not stated
Approaches that could also have been used
  • Pooled effects were combined using a restricted maximum likelihood random-effects meta-analysis.
    Could also: A DerSimonian-Laird random-effects model, or a fixed-effect inverse-variance model, could also be used to pool study-level effects. — Comparing estimators (REML vs DerSimonian-Laird vs fixed-effect) can show how robust pooled estimates and confidence intervals are to the choice of between-study variance estimator.
  • Between-study variability was handled within the random-effects framework.
    Could also: Heterogeneity statistics such as I² or Cochran's Q, and prediction intervals alongside confidence intervals, could also be reported. — These quantify how much effects vary across studies and how generalizable a pooled estimate is, complementing the pooled point estimate.
  • Multiple-comparison control was reported via FDR thresholds (e.g., <1%, <0.1%).
    Could also: Naming the specific FDR procedure (e.g., Benjamini-Hochberg or Storey's q-value) could also be done. — Stating the exact procedure makes the multiplicity adjustment fully reproducible and lets readers map thresholds to expected false-discovery proportions.
  • Inter-study structure was summarized with PCA, chord plots and fold-change correlation matrices.
    Could also: Hierarchical clustering with bootstrap support, or model-based batch/covariate adjustment (e.g., including platform or study as a random factor), could also be used. — These approaches can quantify clustering confidence and separate biological signal from platform/study-level technical variation across heterogeneous datasets.
  • Top genes were validated by qPCR and assessed via correlation with MetaMEx estimates.
    Could also: Concordance metrics such as Lin's concordance correlation or Bland-Altman agreement could also accompany the correlation. — Agreement-based metrics capture systematic offsets in magnitude, not just rank/linear association, when comparing platforms.
  • Subgroups were defined as Healthy versus Metabolically Impaired due to study heterogeneity.
    Could also: Meta-regression or formal subgroup interaction tests on covariates (e.g., BMI, age, sex, muscle type) could also be applied. — Meta-regression models can directly estimate how moderators relate to the transcriptional response rather than relying on dichotomized grouping.
Software: MetaMEx (custom online tool/database, www.metamex.eu)

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
451
Impact: very high
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.

GSE27285 GEO in Methods (http://purl.org/orb/Methods)
also used by 1 paper:
GSE43856 GEO in Methods (http://purl.org/orb/Methods)
also used by 1 paper:
GSE42507 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE44818 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE5792 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE74194 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
RRID:AB_2756818 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.

Figures / tables: Fig.3Fig.2
C1
Reported
NR4A3 logFC acute aerobic 2.99 (FDR 2.0E-07)
Reproduced
3.99 (FDR_bonf 1.0E-23), k=49
partial
C2
Reported
NR4A3 logFC acute resistance 2.95 (FDR 8.2E-15)
Reproduced
2.45 (FDR_bonf 3.5E-05), k=35
partial
C3
Reported
NR4A3 logFC inactivity -0.32 (FDR 9.4E-02)
Reproduced
-0.53 (raw p 1.1E-04), k=16
partial
C4
Reported
PPARGC1A acute aerobic 2.3-fold [1.6,3.5]
Reproduced
2.24-fold [1.70,2.94]
within tolerance
C5
Reported
PPARGC1A acute resistance 1.8-fold [1.6,2.2]
Reproduced
1.60-fold [1.29,1.99]
partial
C6
Reported
PPARGC1A inactivity 25% decrease
Reproduced
24.6% decrease
within tolerance
C7
Reported
sig genes acute aerobic FDR<0.1% = 897
Reproduced
949 bonf / 3162 BH
partial
C8
Reported
sig genes acute resistance FDR<0.1% = 2404
Reproduced
2354 BH / 687 bonf
partial
C9
Reported
sig genes inactivity FDR<0.1% = 1576
Reproduced
1460 BH / 507 bonf
partial
C10
Reported
sig genes aerobic training FDR<0.1% = 82
Reproduced
32 bonf / 175 BH
partial
C11
Reported
sig genes resistance training FDR<0.1% = 2049
Reproduced
1597 BH / 297 bonf
partial
C12
Reported
overlap acute aerobic & resistance = 360
Reproduced
185 bonf / 935 BH
did not match
C13
Reported
overlap aerobic & resistance training = 25
Reproduced
8 bonf / 76 BH
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 52/100

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: Q6 · Severity 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 +8

The MetaMEx meta-analysis pipeline was ported line-for-line (metafor::rma REML on shipped per-study stats) and reproduces robustly: PPARGC1A lands at 2.24 vs 2.3-fold (acute aerobic) and -24.6% vs -25% (inactivity), and NR4A3 reproduces in direction in all conditions. The deviations (NR4A3 3.99 vs 2.99; overlaps 360 vs 185/935; counts off ~2-22%) sit on the input/data-availability side: the repo only ships the larger v3.2208 (2022) database whereas the paper used v1 (2020), which is undeposited, plus a BH-vs-Bonferroni FDR ambiguity. This is a solid, explainable reproduction with no fabrication signal — not an authors' defect — limited mainly by version drift and the original version not being archived.

🤝
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.

161.6 k
tokens (I/O) · 9.3 M incl. cache
19 min
runtime · 0.31 CPU-h
5.8 GB
peak RAM
1
HPC jobs
hummel
machine