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

Contribution of retrotransposition to developmental disorders.

Nat Commun · 2019
L1 79/100 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) 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score 0
✓ What held up
  • Reported values were directly comparable
  • 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)
  • 🟡Could not use the authors’ exact input data
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
79/100
Reproducibility score
0.3 SD above mean
vs. all fields · 1173 studies
🎯 Scores higher than 55% of all assessed papers rank 514 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 — YES, and reproduced 1:1 on the parts that are public. NOTE: the scaffold's auto-enriched links were both false positives (code 'PacBio/unanimity' and data 'GSE74432' are unrelated); the real artifacts are the authors' own repos github.com/eugenegardner/MEIPaper (@0e709a3) + Retrogene (@d65b170), with primary data in EGA controlled access (EGAS00001000775). The raw exomes/BAMs are restricted, so MELT/Retrogene variant CALLING could not be re-run (a real, declared blocker). HOWEVER the MEIPaper repo ships the complete downstream R analysis together with its input summary tables, and that pipeline reproduces the paper's headline numbers EXACTLY on «our HPC»: all six Table 1 per-individual MEI/PPG means+SD (Alu 23.61±4.25 vs 23.6±4.2, total 26.57±4.68 vs 26.6±4.7, etc.), both Table 2 germline mutation rates (1.21e-11 DDD, 1.40e-11 1KGP), the '1 new MEI per 12-14 births' genome-wide rate (12.42-14.39), and the coin-flip P=1.7e-5 (~10^-5 as stated). Supplementary statistics (1KGP downsampling z-tests, mosaic Wilcoxon, PPG-vs-Zhang regression R2=0.64, pLI Fisher, strand chisq) also recomputed deterministically and recorded. The 1-in-2434 (0.04%) diagnostic figure is arithmetically confirmed (4/9738=1/2434.5=0.041%). NOT ATTEMPTED (out of scope): MELT/Retrogene calling from restricted EGA data, the 9-de-novo and 4-causative counts (need restricted data + manual clinical review), wet-lab Figure 3 PCR/blots. Verdict: faithful 1:1 reproduction of the entire shipped (downstream) pipeline; no fabrication signal — every public claim is derivable from the shipped data+code. Honest partial because upstream calling is gated by controlled-access data, not by any repo/doc deficiency.

💻 Code ↗ 🗄 Data: GSE74432

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 79
    assessed: 2026-06-16 ⛓ 62a585393737
✎ 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 tests whether and at what rate de novo retrotransposition (RT) events—mobile element insertions (MEIs) and processed pseudogenes (PPGs)—contribute to severe developmental disorders, and seeks to estimate genome-wide germline ME mutation rate and selective constraint.

Core claims
  • De novo retrotransposition events cause a small but detectable fraction of severe developmental disorders, with 4 of 9 de novo MEIs deemed likely causative (~0.04%). finding
  • Coding MEIs are under strong purifying selection equivalent to that of nonsense/essential splice-site (protein-truncating) SNVs. finding
  • MELT can reliably ascertain MEIs from whole-exome sequencing data, validated against matched WGS. method
  • A bespoke tool detects processed pseudogenes genome-wide from WES via cross-hybridization of donor-gene exome baits. method
  • PPG donor genes are significantly enriched for loss-of-function intolerant genes (high pLI) and highly expressed genes, consistent with prior retroduplication patterns. finding
  • Intronic MEIs show no detectable selective constraint, behaving similarly to synonymous SNVs. finding
  • A comprehensive catalogue of RT-derived variation (1129 MEIs, 576 polymorphic PPGs) across 9738 DDD trios serves as a resource and framework for evolutionary and disease studies. resource
  • Most de novo RT events phaseable in this study are of paternal origin, consistent with other de novo variant classes. finding
Experimental setups
Assay System Perturbation Readout Platform
Whole-exome sequencing (trio) with MEI calling via MELT 9738 DDD trios (n=28,132 individuals), human none Alu/L1/SVA mobile element insertions in/adjacent to WES bait regions MELT (mobile element locator tool)
Processed pseudogene detection from WES DDD cohort, human exomes none presence/absence and allele counts of PPGs via discordant read pairs and split reads bespoke PPG detection tool
Matched whole-genome sequencing concordance comparison 90 overlapping DDD individuals, human none MEI genotype concordance between WES and WGS
Long-PCR coupled to long-read sequencing (phasing) DDD parent-proband samples, human gDNA none phasing of false-negative Alu variants to nearby SNVs / mosaicism assessment
PCR validation (and capillary sequencing) proband/paternal/maternal gDNA, human none de novo status confirmation of MEIs and PPGs
Gene functional annotation and expression analysis PPG donor genes across 30 GTEx tissue types, human none functional enrichment (DAVID) and tissue expression of donor genes GTEx / DAVID
Key results
  • Identified 9 de novo MEIs (7 Alu, 2 L1) and 2 de novo gene retroduplications across 9738 trios 11 total de novo RT events
  • 4 de novo MEIs disrupted coding sequence of known DD genes (SETD5, NSD1, MEF2C, ARID2) and were likely causative 0.04% of probands
  • Exonic MEIs show selective constraint indistinguishable from nonsense and essential splice-site SNVs
  • WES MEI genotype concordance with matched WGS 94.46% overall (93.93% Alu, 97.29% L1, 98.25% SVA)
  • Re-identified MEI genotypes detectable in WGS within WES bait regions 1450 genotypes, 84.5%
  • PPG donor genes enriched for high-pLI (LoF-intolerant) genes vs all protein-coding genes 25.3% vs 17.6%
  • Rate of de novo MEI events approximately one per 1000 patient exomes ~1/1000
  • Each human genome harbors ~one MEI directly impacting protein-coding sequence; MEIs make up 0.6% of coding PTVs 0.76 ± 0.62 SD per individual
Key statistics
  • other 94.46% (93.93% Alu, 97.29% L1, 98.25% SVA) (WES vs WGS MEI genotype concordance, ≥10× coverage)
  • count 1129 MEIs and 576 polymorphic PPGs; 33.0 ± 5.0 SD variants per exome (total RT variants discovered in DDD trios)
  • pvalue Fisher's p = 2.6 × 10–4 (PPG donor genes enriched for high-pLI (>0.9) genes (25.3% vs 17.6%))
  • correlation r2 = 0.64 (PPG relative allele counts vs Zhang et al. 1KGP assessment)
  • other r2 = 0.64 / DAVID enrichment score 8.82 (ribosomal/translational gene functional cluster among PPG donor genes)
  • pvalue χ2 p = 0.083 (paternal-origin bias of de novo RT events (3 phased, all paternal))
  • count 84.5% (1450 genotypes) (proportion of WGS-identifiable MEI genotypes re-identified in WES bait regions)
  • other 1 incorrect candidate de novo per 295 patients (1 false-positive per 649; 1 false-negative per 541) (false finding rate of the RT assessment pipeline)

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 population/statistical genomics study that systematically ascertains retrotransposition (RT)-derived variants (mobile element insertions and processed pseudogenes) from 9738 whole-exome-sequenced parent–child trios (28,132 individuals) in the DDD cohort. The analysis is largely descriptive and enumerative—reporting concordance/validation rates, per-individual variant counts summarized as mean ± SD, allele-frequency and Poisson-distribution fits—and uses categorical/count-based tests (chi-square, Fisher's exact, Wilcoxon rank-sum) plus Poisson-based simulation to compare observed versus expected de novo counts and to assess selective constraint. Results are reported with point estimates, exact or threshold p-values, and 95% confidence intervals on key proportions.

Replicationbiological Sample sizeCohort of 9738 trios (28,132 individuals); constraint analysis restricted to 17,032 unaffected parents to avoid relatedness and ascertainment bias; expected de novo counts derived from 100 Poisson simulations; no formal a priori power calculation described GroupsRT/MEI variant classes vs SNV classes; observed vs expected de novo counts; PPG donor genes vs all genes; cases vs population reference (1KGP, GTEx) Pairingna Randomization/blindingna DispersionSD Exact p-valuesyes Effect sizesyes Confidence intervalsyes Multiplicity correctionnone stated
Statistical tests used
Test Applied to n Assumptions
Wilcoxon rank-sum (Mann–Whitney) test comparison of PPG donor-gene expression vs non-retroposed genes across GTEx tissues, and testis/ovary vs other tissues 30 GTEx tissue types (28 compared for ovary/testis) not stated
Fisher's exact test enrichment of high-pLI (>0.9) genes among PPG donor genes (25.3%) vs all protein-coding genes (17.6%) not stated
Chi-square (χ²) test intronic MEI sense/antisense orientation bias; paternal vs maternal parental origin of de novo events 3 of 11 de novo events phased (parental origin) not stated
Poisson-distribution test based on simulation observed vs expected de novo MEI counts in exons and enhancers for all/high-pLI/MA-DDG2P gene sets (Fig. 4) 100 simulations using neutral mutation rate (1.2 × 10⁻¹¹ μ) stated (Poisson model assumed; counts approximate Poisson)
Correlation (r²) between data sets PPG allele counts in this study vs Zhang et al. 1KGP assessment (r²=0.64); PPG vs WGS validation na
Genotype concordance / proportion estimation WES vs matched WGS MEI calls (94.46% overall concordance); singleton and pLI proportions for constraint (Fig. 2b) 90 overlapping WGS individuals; 17,032 unaffected parents for constraint na
Approaches that could also have been used
  • Per-individual variant counts were summarized as mean ± SD.
    Could also: The same spread could also be conveyed with a 95% confidence interval or, given the count/Poisson nature of the data, an interquartile range or the Poisson rate with its interval. — A CI communicates precision of the mean estimate, and an IQR/Poisson interval can be a natural complement for skewed count data; all are standard ways to convey dispersion.
  • Multiple per-tissue Wilcoxon rank-sum comparisons (across ~30 GTEx tissues) were evaluated against a fixed p<1×10⁻³ threshold.
    Could also: A formal multiplicity adjustment such as Benjamini–Hochberg FDR or Bonferroni across the tissue family could also be applied alongside the threshold. — An explicit correction makes the family-wise or false-discovery control transparent when many tissues are tested, complementing the stringent fixed cutoff already used.
  • Observed vs expected de novo counts were assessed using a Poisson model based on a neutral mutation rate and 100 simulations.
    Could also: An exact Poisson test, a negative binomial model (to allow overdispersion), or a larger number of simulations could also be used. — A negative binomial or exact approach can accommodate possible overdispersion, and more simulation replicates can tighten the empirical null; these can add robustness for rare-count inference.
  • Categorical comparisons (orientation bias, parental origin) were evaluated with chi-square tests, some on small counts (e.g. 3 of 11 events phased).
    Could also: Fisher's exact test (or an exact binomial test) could also be used for these small-count comparisons. — Exact tests do not rely on large-sample chi-square approximations and are often preferred when expected cell counts are small.
  • Agreement between data sets (e.g. PPG allele counts vs 1KGP) was summarized with an r² value.
    Could also: A concordance correlation coefficient, Spearman's rank correlation, or a Bland–Altman-style agreement plot could also be reported. — Rank-based or agreement-specific metrics can characterize monotonic association and systematic offsets, complementing r² for cross-platform comparisons.
  • Constraint was assessed via two summary proportions (singletons and fraction in high-pLI genes) with 95% CIs.
    Could also: A logistic regression or other model adjusting for covariates (e.g. variant length, coverage) could also estimate constraint effects. — A modeling approach can jointly account for potential confounders and yield adjusted effect estimates, complementing the descriptive proportion comparison.
Software: MELT (Mobile Element Locator Tool) · Bespoke/custom PPG detection tool · DAVID (functional annotation)

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
60
Impact: 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.

ENSG00000006576 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000081189 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000092820 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000114346 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000163751 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000165671 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000168137 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000169504 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000174579 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
ENSG00000189079 Ensembl in Results (http://purl.org/orb/Results)
no other assessed paper uses this yet
GSE74432 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-31604926

Paper: Gardner EJ et al. "Contribution of retrotransposition to developmental disorders." Nat Commun 10, 4630 (2019). PMCID PMC6789007. DOI 10.1038/s41467-019-12520-y.

Correction to auto-enriched metadata (IMPORTANT)

The scaffold's enrichment auto-filled the WRONG artifacts:

  • code_url = github.com/PacificBiosciences/unanimityFALSE POSITIVE (PacBio CCS consensus tool, unrelated). The real code is the authors' own repos (below).
  • data_accession = GSE74432 (GEO)FALSE POSITIVE as the primary dataset. GSE74432 is only a methylation reference used as a side input. The paper's primary data are DDD exomes in EGA (controlled access).

Real artifacts (from the paper's Code/Data availability)

  • Code (authors' own):
    • github.com/eugenegardner/MEIPaper @ commit 0e709a30a0f8c8667600f2af7495f14f435524f9 — R-markdown notebooks (Figure1/2/4 + Supplement) that regenerate all figures & statistics derived in R, shipping their own RawData/ (small text files).
    • github.com/eugenegardner/Retrogene @ commit d65b17039390c8bd73dc7fd7b325c846618f5c06 — Java pseudogene/PPG discovery pipeline (operates on BAMs).
  • Primary data (raw): EGA controlled access — study EGAS00001000775, datasets EGAD00001004390 (WES), EGAD00001004586 (RT calls + WGS). dbGaP/EGA controlled → NOT publicly obtainable.

In scope (reproducible from shipped code + shipped data) — ATTEMPTED

The MEIPaper repo ships the post-calling R analysis with the exact small input tables it consumes. This is the "pipeline-derived" downstream analysis. Pipeline = R (data.table/ggplot2/cowplot) over the shipped MELT/Retrogene output summaries.

Result Where in paper Rmd / data
Mean ascertained MEIs per individual exome (Alu/L1/SVA/PPG/total) Table 1 Figure1.Rmd ← *.totals.di.txt
Germline MEI mutation rate (DDD μ & 1KGP, ~1.2–1.4×10⁻¹¹ /bp/gen) Table 2 / Results Figure4.Rmd ← hardcoded site totals
Genome-wide rate: ~1 new MEI per 12–14 births Results Figure4.Rmd
Expected # de novo MEIs in 9738 individuals Results Figure4.Rmd
"Coin-flip" P that 4/6 exonic de novos hit MA-DDG2P (~10⁻⁵) Results/Methods Figure4.Rmd
de novo enrichment Poisson P-values (Fig 4) Fig 4 Figure4.Rmd ← de_novo.test.txt
1KGP downsampling significance (Alu 318 / L1 81 / SVA 26 vs DDD) Supp Fig 4 Supplement.Rmd ← *_random.txt
Mosaic/VAF Wilcoxon tests (parents vs probands vs hets) Supp Fig 2 Supplement.Rmd ← genodepth.di.txt
PPG vs Zhang/Gerstein allele-count regression r² Supp Fig 5 Supplement.Rmd ← gerstein.txt, Retrogenes.di.txt
PPG tissue-expression / pLI Fisher test Supp Fig 6 Supplement.Rmd ← GTEx.expr.txt, ExacPLI.txt
Intronic MEI strand-bias chi-square (Alu/L1/SVA) Supp Fig 8 Supplement.Rmd ← strands.txt
MEI allele-frequency spectrum & length distributions Fig 1F/1G Figure1.Rmd ← *.frq, *.len
Exon/intron pLI enrichment (Fig 2) Fig 2 Figure2.Rmd

Out of scope (cannot reproduce from public artifacts) — NOT attempted, with reason

Step Reason
MELT MEI calling from raw exomes Input = EGA controlled-access DDD WES (EGAD00001004390). Not obtainable.
Retrogene/PPG calling from raw BAMs Same restricted BAMs; Java tool present but no input data.
Wet-lab validation (PCR, capillary, Figure 3 blots) Manual/experimental; not a pipeline.
Clinical pathogenicity calls ("4 likely causative") Manual clinician review; not computational.

Reproduction strategy

Raw-data re-calling is blocked by controlled access (a legitimate partial blocker), but the downstream statistical pipeline is fully shipped (code + its input summary tables). We rerun the R notebooks on «our HPC» and compare the regenerated numbers 1:1 against Table 1, Table 2, and the Results/Supplement text. This is an "own-code on shipped data" reproduction — honest partial: downstream analy

Figures / tables: Table
C1_alu_per_indiv
Reported
23.6 ± 4.2
Reproduced
23.6138 ± 4.2464
exact
C2_l1_per_indiv
Reported
2.8 ± 1.5
Reproduced
2.7700 ± 1.4647
exact
C3_sva_per_indiv
Reported
0.2 ± 0.5
Reproduced
0.1836 ± 0.4683
exact
C4_totalmei_per_indiv
Reported
26.6 ± 4.7
Reproduced
26.5673 ± 4.6813
exact
C5_ppg_per_indiv
Reported
6.6 ± 1.7
Reproduced
6.5501 ± 1.7059
exact
C6_combined_per_indiv
Reported
33.0 ± 5.0
Reproduced
33.0267 ± 4.9833
exact
C7_mu_ddd
Reported
1.2e-11 /bp/gen
Reproduced
1.21e-11
exact
C8_mu_1kgp
Reported
1.4e-11 /bp/gen
Reproduced
1.40e-11
exact
C9_births_per_mei
Reported
12-14 births per new MEI
Reproduced
12.42-14.39 births
exact
C10_coinflip_p
Reported
~1e-5
Reproduced
1.7e-5
within tolerance
C11_downsample_alu_p
Reported
Supp Fig 4 (Alu 318)
Reproduced
p=4.96e-2
partial
C12_downsample_l1_p
Reported
Supp Fig 4 (L1 81)
Reproduced
p=3.05e-6
partial
C13_downsample_sva_p
Reported
Supp Fig 4 (SVA 26)
Reproduced
p=8.05e-3
partial
C14_strand_chisq
Reported
Supp Fig 8 strand bias
Reproduced
ALU 0.416 / L1 0.931 / SVA 0.543
partial
C15_ppg_gerstein_r2
Reported
Supp Fig 5 correlation
Reproduced
R2=0.642, p<2.2e-16
partial
C16_denovo_count
Reported
9 de novo MEs / 9738 trios
Reproduced
not recomputed (restricted EGA exomes); repo input 6 exonic + 3 intronic = 9, internally consistent
partial
C17_diagnostic_fraction
Reported
1/2434 (0.04%)
Reproduced
4/9738 = 1/2434.5 = 0.041% (arithmetic confirmed)
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 79/100

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) 🟡
Concordant (toward reproduced)
Code + data deposited & functional
-2 pts
From: Data & code availability Available & functional
Total score 0

On everything testable this is a faithful 1:1 reproduction: all six Table 1 per-individual means±SD and both Table 2 germline rates (1.21e-11 / 1.40e-11) match to printed precision, the 12–14 births/MEI range and coin-flip P (~1e-5) reproduce, with deviations only at the rounding/stochastic level and no fabrication signal. The only gaps sit on the data-availability side, not the authors' side: primary exomes/BAMs are EGA controlled-access, so MELT/Retrogene calling, the 9 de-novo count (C16), and the manually-curated numerator behind the 1/2434 diagnostic fraction (C17) could not be independently re-derived. Net: exact where reproducible, honestly partial where gated by restricted data — overall solid but not a clean full-coverage green.

🤝
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.

141.8 k
tokens (I/O) · 13.4 M incl. cache
20 min
runtime · 0.02 CPU-h
1.7 GB
peak RAM
2 (1 failed)
HPC jobs
hummel
machine