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

Single-cell multiomics profiling reveals heterogeneous transcriptional programs and microenvironment in DSRCTs.

Cell Rep Med · 2024
L1 76/100 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: Q5 · Derivability / plausibility 🟡
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 +6
✓ What held up
  • The central claim held under reproduction
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
  • 🟡Overall, the reproduction showed a material discrepancy
How its reproducibility compares
76/100
Reproducibility score
at the mean
vs. all fields · 1173 studies
🎯 Scores higher than 48% of all assessed papers rank 586 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 a 1:1-in-distribution reproduction of the Int_sc 3' scRNA-seq pipeline; NOT exact-to-the-digit on the headline cell count. The paper reports NO original code ('available upon request'); the brief's code_url is the third-party Harmony tool (Harmony v1.0, cited in STAR Methods). Per P16 we reproduced by applying harmonypy (Python port of Harmony) + a standard scanpy QC/clustering pipeline to the paper's OWN GEO count matrices (GSE263523, 11 3' scRNA-seq libraries GSM8193966-76), all compute on «our HPC»/«infra» (final «job»; earlier jobs fixed conda-on-«infra» env + a harmonypy obsm-orientation bug). RESULTS: (C1) median genes/cell = 1,967 vs reported 1,971 -- essentially EXACT (0.2%); the total high-quality cell count is NOT reproducible to the digit because authors' QC/cell-calling thresholds are unspecified and GEO ships mostly RAW matrices -- we bracketed it [40,276 @floor500 .. 94,706 @floor200] and 51,671 falls inside. (C2) malignant fraction 85.0% vs 86.7% -- within 1.7 pts. (C3) microenvironment composition reproduces strikingly: identical rank order CAF>myeloid>lymphoid>mesothelial>endothelial and per-population proportions within ~1.5 pts (endothelial 9.5% identical). NO fabrication indicated -- the paper's 5 stromal counts sum exactly to the non-malignant total (6,890) and our independent re-run recovers the same structure. NOT ATTEMPTED (hard ~20%): exact cell count (unspecified QC), snMultiome/snATAC peak+WNN analysis, Visium spatial deconvolution, inferCNV-based malignant calling, SCENIC/Hotspot regulons, CellPhoneDB interactions, bulk RNA-seq DE, ChIP/CUT&RUN, and all wet-lab/functional assays -- separate pipelines/assays out of scope.

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 76
    assessed: 2026-06-14 ⛓ a893971c629a
✎ I am an author of this paper

Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.

Reason for the rerun

We email you a confirmation link first. The rerun is an objective re-measurement — it cannot change the verdict in your favour, only ask us to look again.

Provenance — full disclosure

When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.

Reproduced
2026-06-14
Rubric version
v1.0
Assessed by
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-15
no human curator yet
Last updated
2026-08-05

Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.

Deep full-text extraction

Model: opus
Founding hypothesis

Does DSRCT intratumor heterogeneity and polyphenotypic differentiation arise from cell-intrinsic transcriptional plasticity linked to variable EWSR1::WT1 activity and/or cell-extrinsic microenvironmental cues, despite the tumor being driven by a single EWSR1::WT1 oncogenic driver?

Core claims
  • DSRCT tumor cells cluster into consistent subpopulations with partially overlapping lineage- and metabolism-related transcriptional programs across patients and tumor sites finding
  • DSRCT cell heterogeneity is driven, at least in part, by transcriptional plasticity rather than somatic genetic subclonal evolution mechanism
  • High EWSR1::WT1 DNA-binding activity associates with most lineage-related states, in contrast to glycolytic and profibrotic states finding
  • EWSR1::WT1 binding site variability may drive distinct lineage-related transcriptional programs, supporting cell-intrinsic plasticity mechanism
  • Glycolytic and profibrotic states localize within hypoxic niches at the periphery of tumor cell islets, implicating cell-extrinsic microenvironmental cues finding
  • A single-cell transcriptomics-derived epithelial signature is associated with improved patient survival across two independent cohorts finding
  • EWSR1::WT1 binds regulatory regions of lineage-related and developmental/metabolic genes and is enriched at EGR1-related [GGA]-repeat consensus motifs mechanism
  • Copy number variations are homogeneous across DSRCT tumor cell clusters and samples, consistent with a quiet genomic profile finding
Experimental setups
Assay System Perturbation Readout Platform
3' single-cell RNA-seq 10 fresh human DSRCT samples plus one juxtatumoral peritoneal sample none single-cell gene expression / cell type composition and tumor cell clustering 10x Genomics Chromium 3'
single-nucleus Multiome (snRNA-seq + snATAC-seq) one fresh human DSRCT tumor sample none single-nucleus gene expression and chromatin accessibility 10x Genomics snMultiome
bulk whole exome sequencing (WES) 10 single-cell-matched DSRCT samples none somatic mutations and copy number variations
bulk RNA-seq 29 archived frozen DSRCT samples none gene expression; CIBERSORTx cell-type deconvolution
ChIP-seq JN-DSRCT-1 cell line none (anti-WT1 C-terminus to pull down EWSR1::WT1) EWSR1::WT1 genome-wide binding peaks and motifs WT1 C-terminus antibody
EWSR1::WT1 silencing with chromatin profiling (in vitro modeling) DSRCT cell line (in vitro) EWSR1::WT1 knockdown/silencing chromatin landscape and transcriptome upon EWSR1::WT1 modulation
Immunohistochemistry DSRCT patient tissue samples none protein staining for WT1, AE1/AE3, Desmin, CD56, THY1, CD68/CD163, CD3
custom EWSR1::WT1 transcript amplification assay 10x-derived barcoded cDNAs from DSRCT single cells none EWSR1::WT1 transcript expression level per cell
Key results
  • After QC, 51,671 high-quality cells were identified including 3,063 from the juxtatumoral peritoneal sample 51,671 cells
  • Malignant cells represented 92% of detected cells after excluding the juxtatumoral sample, with CAFs <5% and myeloid cells <4% 92% tumor; CAFs <5%; myeloid <4%
  • 13 DSRCT tumor cell clusters delineated: six lineage-, two metabolic-, and five pseudostate-related 13 clusters
  • Rare identified CNVs (most frequent: chromosome 5 gain) were highly consistent across single-cell clusters within each sample
  • EWSR1::WT1 transcript expression variation was limited across cancer cells and did not correlate with DGE-derived clusters
  • EWSR1::WT1 ChIP-seq identified consensus peak regions enriched in developmental, stem cell, and fatty acid metabolic processes 8,782 and 4,507 consensus peaks (replicates); 1,587 shared
  • GGAGGA 6-mers predominant within EWSR1::WT1 peaks; top de novo motifs matched WT1+/-KTS, EGR1/2/3, and ZNF263 8 de novo motifs
  • Most tumor cell clusters predicted less differentiated than non-tumor cells (except ribosomal/lncRNA pseudostates); no consistent transcriptional trajectory found
Key statistics
  • count 51,671 high-quality cells (Int_sc integrated scRNA-seq dataset after QC)
  • count 44,781 tumor cells (DSRCT malignant cells identified)
  • count 3,063 cells (cells from juxtatumoral peritoneal sample)
  • count median 1,971 expressed genes per cell (Int_sc dataset cell quality)
  • other mean FRIP-score of 0.81% (EWSR1::WT1 ChIP-seq coverage quality)
  • other relative strand correlation 0.73 and normalized strand coefficient 1.02 (ChIP-seq quality metrics)
  • count 8,806 and 4,541 peaks for two WT1 ChIP replicates (EWSR1::WT1 ChIP-seq peak calling)
  • count 2,360 CAFs; 1,966 myeloid; 1,126 lymphoid; 784 mesothelial; 654 endothelial (non-malignant cell counts in Int_sc dataset)

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.

This study characterized DSRCT heterogeneity using single-cell RNA-sequencing (51,671 cells from 12 samples across 5 patients), single-nucleus multiome (snRNA-seq + snATAC-seq), bulk RNA-seq (n=29), bulk whole-exome sequencing, ChIP-seq, and spatial transcriptomics. Core analytical strategies were unsupervised gene expression clustering with differential gene expression (DGE) analysis to define cell subpopulations, and gene set enrichment analysis (GSEA) with Gene Ontology terms to characterize cluster-level pathway activity. Supporting analyses included CIBERSORTx deconvolution of bulk RNA-seq, copy number inference, RNA velocity, stemness scoring, and evaluation of a scRNA-seq-derived epithelial signature for prognostic significance in independent cohorts.

Replicationbiological Sample size12 samples from 5 patients for single-cell assays (10 fresh tumor + 1 juxtatumoral by 3' scRNA-seq; 1 by snMultiome); 29 archived frozen samples for bulk RNA-seq; 10 matched samples for WES; no formal power calculation stated — sample size determined by patient availability Groups13 tumor cell clusters vs. each other; tumor vs. non-malignant cell types; DSRCT vs. other sarcoma histotypes (DGE); bulk RNA-seq samples for deconvolution; pre- vs. post-chemotherapy (GR4 vs. GR4_PC, one patient) Pairingmixed Randomization/blindingnot stated Dispersionnone Exact p-valuesno Effect sizesno Confidence intervalsno Multiplicity correctionnot stated for DGE; GO GSEA applied p < 0.05 threshold then ranking by gene ratio
Statistical tests used
Test Applied to n Assumptions
Differential gene expression (DGE) analysis — specific statistical method not named in excerpt (likely Wilcoxon rank-sum or Seurat default); used to derive top 50 DEGs per cluster Identification of marker genes for 13 tumor cell clusters and 6 non-malignant cell types in Int_sc dataset; DSRCT vs. other sarcoma histotypes (Figure 2B, Figure 1G) 44,781 tumor cells; 51,671 total cells across 12 samples not stated
Gene set enrichment analysis (GSEA) with Gene Ontology (GO) terms; threshold p < 0.05; top 5 enriched selected then ranked by gene ratio, top 3 displayed Pathway enrichment per Int_sc tumor cell cluster (Figure 2C); EWSR1::WT1 ChIP-seq target gene enrichment (Figure S2C) not stated
CIBERSORTx reference-based cell-type deconvolution Estimation of cell-type proportions across 29 bulk RNA-seq DSRCT samples (Figure 1I) 29 bulk RNA-seq samples na
HotSpot coexpressed gene module analysis Identification of correlated gene expression modules across DSRCT tumor cells (Figure 2D) 44,781 tumor cells not stated
CytoTRACE and single-cell entropy (differentiation/stemness scoring) Stemness assessment across Int_sc clusters relative to non-tumor cells (Figures S1C, S1D) 51,671 cells not stated
ChIP-seq peak calling and differential binding analysis comparing WT1 ChIP to isotype control; quality assessed by FRIP score, RSC, and NSC; peak intersection across two replicates EWSR1::WT1 binding target identification in JN-DSRCT-1 cell line (Figures S2A–S2B) 2 ChIP replicates; 1 cell line not stated
Copy number variation (CNV) inference from scRNA-seq (method not named in excerpt); compared to matched bulk WES Genetic heterogeneity assessment across tumor cell clusters and samples (Figures 2E, 2F) 10 scRNA-seq samples with matched WES not stated
RNA velocity combined with trajectory inference (directed single-cell fate mapping); result was negative — no consistent trajectory identified Transcriptional trajectory analysis across tumor cell clusters (data not shown) not stated
Single-cell label transfer from fetal development reference gene expression atlas Lineage annotation of Int_sc tumor cells against developmental cell types (Figure S1B) not stated
Approaches that could also have been used
  • Cluster marker genes were identified using DGE analysis treating individual cells as observations (specific test not named, likely Wilcoxon rank-sum per Seurat defaults)
    Could also: Pseudobulk differential expression methods (e.g., DESeq2 or edgeR applied to per-sample aggregated counts) could also identify cluster markers while treating patient as the statistical unit — Pseudobulk approaches account for within-patient correlation across cells, which can reduce inflation of test statistics that arises when thousands of cells from the same individual are modeled as independent; they are increasingly recommended as a complement to single-cell-level tests in multi-patient designs
  • GO pathway enrichment was reported using GSEA with a p < 0.05 threshold and secondary ranking by gene ratio; top 3 terms per cluster were displayed
    Could also: Reporting Benjamini-Hochberg FDR-adjusted q-values (rather than a nominal p threshold) across all tested GO terms would also characterize enrichment significance while accounting for the large number of terms tested simultaneously — When hundreds of GO terms are tested per cluster, FDR-adjusted values provide a standard way to identify which enrichments are robust to multiple comparisons; gene-ratio ranking alone does not reflect term size or the breadth of testing
  • Cell-type proportions in bulk RNA-seq were estimated using CIBERSORTx with scRNA-seq-derived reference signatures
    Could also: Additional deconvolution methods such as EPIC, MuSiC, or BayesPrism could also use the same scRNA-seq-derived signatures and would allow cross-method comparison of proportion estimates — Different deconvolution algorithms apply different statistical frameworks and regularization strategies; concordance across methods strengthens confidence in estimated proportions, particularly for rare cell types such as lymphoid and endothelial cells
  • ChIP-seq reproducibility was assessed by intersecting peaks across two replicates and reporting FRIP, RSC, and NSC quality metrics
    Could also: Irreproducibility Discovery Rate (IDR) analysis is an additional standard approach for ChIP-seq replicate assessment that statistically ranks peak consistency based on rank concordance across replicates — IDR provides a probabilistic threshold for retaining reproducible peaks, complementing intersection-based approaches and giving a more quantitative estimate of which peaks are reliably detected across experiments
  • Stemness was quantified as a continuous score per cell using CytoTRACE and single-cell entropy, with cluster-level comparisons shown
    Could also: Reporting score distributions per cluster (e.g., violin or ridge plots with median and IQR) alongside a non-parametric test comparing clusters (e.g., Kruskal-Wallis with post-hoc Dunn correction) could also characterize differentiation heterogeneity quantitatively — Showing distributional spread within clusters and a formal comparison statistic would complement the directional interpretation (less differentiated vs. more differentiated) and characterize variability in stemness within each cluster
  • EWSR1::WT1 mRNA expression variation across tumor cells was assessed by visual inspection of expression correlation with DGE-derived clusters (Figure S1H)
    Could also: A Spearman or Pearson correlation between the EWSR1::WT1 expression score and cluster-level gene module activity scores could also quantify the association between fusion transcript level and transcriptional state — A correlation coefficient with a confidence interval would provide a quantitative, reproducible measure of the association (or lack thereof) between EWSR1::WT1 expression level and cluster identity, supplementing the visual cluster-overlay approach
Software: 10x Genomics Chromium (scRNA-seq/snMultiome platform) · CIBERSORTx · HotSpot · CytoTRACE · GSEA (Gene Set Enrichment Analysis) · UMAP (Uniform Manifold Approximation and Projection)

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
20
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.

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.

CVCL_9W68 Cellosaurus in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE263523 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
GSE81009 GEO in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
P36962 UniProt in Methods (http://purl.org/orb/Methods)
no other assessed paper uses this yet
RRID:AB_10949503 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2040911 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2074844 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2102369 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2335677 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2335684 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_262054 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2626893 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2688012 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2798316 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2864626 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2892832 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_2895679 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:AB_823528 RRID in Article (http://semanticscience.org/resource/SIO_001029)
no other assessed paper uses this yet
RRID:CVCL_9W68 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-38781959

Paper: Henon C, Vibert J, Eychenne T, et al. (2024) Single-cell multiomics profiling reveals heterogeneous transcriptional programs and microenvironment in DSRCTs. Cell Rep Med 5:101582. PMID 38781959 / PMC11228554 / DOI 10.1016/j.xcrm.2024.101582.

Nature of paper: Original research. Single-cell / single-nucleus multiomics of Desmoplastic Small Round Cell Tumors (DSRCT): 3′ scRNA-seq ("Int_sc" dataset), snMultiome (snRNA + snATAC), Visium spatial transcriptomics, bulk RNA-seq, plus wet-lab (ChIP/CUT&RUN, functional assays).

Code / data artifacts

  • Code: The paper states "This paper does not report original code. All custom code used for the analyses was written with existing software ... and is available upon request."No authors' repository. The brief's code_url (github.com/immunogenomics/harmony) is the third-party Harmony batch-integration tool (Korsunsky et al.), cited in the STAR Methods key-resources table as Harmony v1.0. Per P16, applying this third-party tool to the paper's own data is an equally valid reproduction. Other named software: Cell Ranger v3.0.2, Seurat v3.1.4/4.0.4/4.1, R v3.5.1/4.3.0.
  • Data: GEO GSE263523 (SuperSeries), public, count-matrix format. The 3′ scRNA-seq matrices are samples GSM8193966–GSM8193976 (11 tumor + 1 juxtatumoral = 12 libraries) plus the snMultiome RNA matrix (GSM8193977). Spatial (Visium), ChIP/CUT&RUN bigWigs, and bulk RNA-seq are separate subseries. Note: most 3′ matrices are shipped as raw (unfiltered, ~6.8M-barcode whitelist; 20 MB barcode files), a few as filtered (20–30 KB barcode files), so reaching the authors' exact post-QC cell set requires their (unspecified) cell-calling + QC thresholds.

In scope (pipeline-derived, attempt) — the "Int_sc" 3′ scRNA-seq pipeline

id reported claim paper location how to reproduce
C1 "we identified 51,671 high-quality cells ... with a median of 1,971 expressed genes per cell" Results §1, Fig 1 Load the 12 3′ scRNA matrices from GEO, run standard QC + Harmony integration (scanpy/harmonypy = Python port of the same Harmony algorithm); count cells, median genes/cell
C2 "Tumor cells (n = 44,781)" vs non-malignant (6,890) → ~86.7 % malignant Results §1, Table S2A After integration + clustering, split malignant vs non-malignant by canonical markers; compare the fraction (robust to exact QC)
C3 microenvironment composition: CAFs 2,360, myeloid 1,966, lymphoid 1,126, mesothelial 784, endothelial 654 (Σ = 6,890) Results §1, Fig 1F/1G, Table S2A Marker-based annotation of non-malignant clusters (COL1A1/SPARC, C1QA/CD68, NKG7/CD3E, CAV1/MSLN, VWF/PECAM1); recover the populations + relative proportions
C4 12 3′ scRNA-seq libraries incl. one juxtatumoral peritoneal sample (3,063 cells) Results §1, Table S1A Bookkeeping vs GEO sample list; per-sample barcode/cell counts

Out of scope / not attempted (the hard ~20%)

  • Exact 51,671 to the cell. QC thresholds (min genes, max mito %), the doublet-removal step, and the Cell Ranger cell-calling that produced the authors' filtered set are not specified in the paper, and GEO ships mostly raw matrices. We reproduce the headline number with a standard QC pipeline and report the value we obtain + sensitivity, but do not claim a to-the-cell match.
  • snMultiome (snATAC) peak/WNN analysis, Visium spatial deconvolution, SCENIC/ Hotspot regulons, CellPhoneDB interactions, bulk RNA-seq DE, ChIP/CUT&RUN, and all wet-lab/functional results — separate pipelines/assays, not attempted here.
  • CNV-based malignant calling (inferCNV) — the paper labels tumor cells via DSRCT-specific features (EWSR1-WT1 signature); we approximate malignant vs non-malignant by canonical lineage markers, which is sufficient for the composition fr
Figures / tables: Fig 1TableFig 1F
C1
Reported
51,671 high-quality cells; median 1,971 expressed genes/cell
Reproduced
median genes/cell = 1,967; total cells 40,276 (floor min_genes>=500) bracketed by 94,706 (floor>=200), reported 51,671 inside the bracket
within tolerance
C2
Reported
Tumor cells n=44,781 of 51,671 => 86.7% malignant
Reproduced
malignant fraction 85.0% (34,226 malignant / 6,050 non-malignant of 40,276)
within tolerance
C3
Reported
microenvironment: CAF 2,360; myeloid 1,966; lymphoid 1,126; mesothelial 784; endothelial 654
Reproduced
CAF 2,090 (34.5%); myeloid 1,635 (27.0%); lymphoid 1,027 (17.0%); mesothelial 724 (12.0%); endothelial 574 (9.5%) - identical rank order, proportions within ~1.5 pts
within tolerance
C4
Reported
12 3' scRNA-seq libraries (11 tumor + 1 juxtatumoral)
Reproduced
11 3' scRNA-seq libraries on GEO (GSM8193966-76), all processed; per-sample cell counts recorded
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 76/100

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: 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 +6

A solid, in-distribution reproduction of the DSRCT single-cell landscape with no fabrication signal: median genes/cell reproduces to 0.2% (1,967 vs 1,971), the malignant fraction lands within 1.7 pts (85.0% vs 86.7%), and the five-population microenvironment recovers in identical rank order within ~1.5 pts. The deviations are on our-methodology / authors'-incompleteness side — the paper ships no code and omits QC thresholds while GEO provides raw matrices, so the headline 51,671 count can only be bracketed [40,276–94,706] and absolute counts run ~12-22% low. Severity is moderate but direction- and magnitude-preserving, and the central conclusion holds, so overall yellow rather than red.

🤝
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.

181.1 k
tokens (I/O) · 15.6 M incl. cache
32 min
runtime · 0.32 CPU-h
10 GB
peak RAM
4 (1 failed)
HPC jobs
hummel
machine